From Telegraphic Speech to Morphological and Syntactic Iconicity Early Child Speech Development from Polish and English Data
1
2
Ewa Konieczna
From Telegraphic Speech to Morphological and Syntactic Iconicity Early Child Speech Development from Polish and English Data
WYDAWNICTWO UNIWERSYTETU RZESZOWSKIEGO RZESZÓW 2014
3
Recenzował prof. dr hab. BOGDAN SZYMANEK
Opracowanie techniczne KRYSTYNA BARAN Łamanie ANDRZEJ LEWANDOWSKI Projekt okładki AGNIESZKA DZIAMA
© Copyright by Wydawnictwo Uniwersytetu Rzeszowskiego Rzeszów 2014
ISBN 978-83-7996-067-5 1060 WYDAWNICTWO UNIWERSYTETU RZESZOWSKIEGO 35-959 Rzeszów, ul. prof. S. Pigonia 6, tel. 17 872 13 69, tel./faks 17 872 14 26 e-mail:
[email protected]; http://wydawnictwo.ur.edu.pl wydanie I; format B5; ark. wyd. 13,25; ark. druk. 13,375; zlec. red. 83/2014 Druk i oprawa: Drukarnia Uniwersytetu Rzeszowskiego
4
For my parents, Wanda and Stanisław and my daughter Kinia, one of the objects of my study
5
6
Table of contents
Acknowledgements ..................................................................................................................
9
Introduction ............................................................................................................................. 10 C H AP TE R I General overview of first language acquisition theories ....................................................... 14 Introduction ............................................................................................................................... 1.1. Milestones in first language acquisition research ........................................................ 1.1.1. Earliest records of first language acquisition .................................................... 1.1.2. Behaviourism .................................................................................................... 1.1.3. Alternatives to behaviourism ............................................................................ 1.2. Chomsky’s Universal Grammar ................................................................................... 1.3. First grammatical analyses .......................................................................................... 1.4. Theories stemming from the UG approach ................................................................. 1.5. New developments ...................................................................................................... 1.5.1. Minimalism and functional categories .............................................................. 1.5.2. Cognitive theories of first language acquisition ................................................ 1.5.3. Modern cognitive approaches to first language acquisition .............................. Conclusion .................................................................................................................................
14 15 15 15 16 18 23 27 30 30 32 37 46
C H AP TE R I I Omission of content and function words ................................................................................ 47 Introduction ............................................................................................................................... 2.1.Language redundancy .................................................................................................. 2.2. Definition of telegraphic speech .................................................................................. 2.3. Explanation of omissions ............................................................................................ 2.3.1. Omission of functors ......................................................................................... 2.3.2. Omission of contentives .................................................................................... 2.4. Ambiguity ................................................................................................................... 2.4.1. Ambiguity of telegraphic speech ...................................................................... 2.4.2. Ambiguity of sentences deprived of content words .......................................... 2.5. Semantic analysis of sentences devoid of content and function words in Polish and English ........................................................................................................................ Conclusion ..................................................................................................................................
47 48 52 54 55 58 75 77 96 112 117
7
C H AP TE R I I I Morphological and syntactic iconicity in child language ...................................................... 119 Introduction ................................................................................................................................ 3.1. Historical background .................................................................................................. 3.1.1. Arbitrary and motivated signs ............................................................................ 3.1.2. Motivated and arbitrary signs in language ......................................................... 3.2. Morphological iconicity................................................................................................ 3.2.1. Natural Morphology .......................................................................................... 3.2.2. Morphological iconicity in the light of the theory of Natural Morphology ....... 3.2.3. Images in children’s language ........................................................................... 3.2.4. Inflectional diagrams in children’s language ..................................................... 3.2.5. Derivational iconicity ........................................................................................ 3.3. Syntactic iconicity ....................................................................................................... 3.3.1. The principle of quantity .................................................................................... 3.3.2. The principle of sequentiality ............................................................................ 3.3.3. The principle of distance ................................................................................... Conclusion ..................................................................................................................................
119 120 121 123 127 127 131 133 137 164 182 182 188 189 190
Conclusions ............................................................................................................................... 192 References ................................................................................................................................. 203
8
Acknowledgements
First and foremost I would like to express my deepest gratitude to Professor Anna Malicka-Kleparska, the supervisor of my PhD dissertation, who has devoted much of her time carefully examining every page of this work and providing me with her professional guidance and valuable support, without which the present shape of this book would never have been possible. Likewise, my sincere thanks go to the late Professor Igor Burkhanov of the University of Gdańsk and Professor Bogusław Marek of the John Paul II Catholic University of Lublin, who, as reviewers of my PhD thesis suggested certain improvements and alterations that have been adopted in the present book. Moreover, I would like to express my gratitude to Professor Pavol Štekauer of the University of Rzeszów and Professor Ruth Berman of the University of TelAviv, who have read substantial portions of this work and whose advice and assistance have been a great help in completing this project. Let me also express my appreciation to Professor Robert Looby, the language consultant, who contributed to the language quality of my work. Finally, I would like to express my gratitude to Professor Bogdan Szymanek for all his remarks and critical comments on the final version of this manuscript. Naturally, none of these people can be blamed for the deficiencies of this study as responsibility for all the errors, misconceptions and blunders remains entirely my own.
9
Introduction
Child: Want other one spoon, Daddy. Father: You mean, you want the other spoon. Child: Yes, I want other one spoon, please Daddy. Father: Can you say ‘the other spoon’? Child: Other…one…spoon. Father: Say ‘other’. Child: Other. Father: ‘Spoon’ Child: Spoon Father: ‘Other spoon’ Child: Other…spoon. Now give me other one spoon? Martin Braine (1971)
The process of first language acquisition is a remarkable phenomenon. By the time a child turns five or six, he or she has become an extremely refined language-user, with a highly developed communicative system that no other creature comes close to matching. The development of the child’s linguistic knowledge is characterised by several features: it proceeds in stages, at great speed and, generally, without any overt instruction. The outcome of each individual acquisition process is the mastery of a highly sophisticated linguistic system, consisting of several subsystems: phonological, morphological, and syntactic. The speed at which children acquire the intricate system of their native language has made the study of language acquisition one of the most fascinating, important and complicated branches of linguistics in recent years. The very nature of the first language acquisition process has aroused numerous controversies among psychologists, linguists and psycholinguists, trying to account for the way in which children master their native language on various grounds. The present book has been based on the doctoral dissertation entitled Early Stages of Morphological and Syntactic Development in Child Speech from Polish and English Data – Longitudinal Study defended at the John Paul II Catholic University of Lublin in 2006. Since then its substantial parts have been published in journals, such as Word Structure, SKASE, or Odisea: Revista de Estudios Inglese as well as post-conference proceedings, e.g. Language, Literature, Culture and Beyond. However, what should be emphasised at this 10
point is the fact that since then a lot of new publications on the first language acquisition have appeared, which have shed a new light on the issues discussed in this publication. Therefore, in no way does the present book arrogate to itself the right to represent the latest outcomes of the child language research. The aim of this work is to venture on a contrastive study of the selected morphological and syntactic mechanisms relied on in the acquisition of English and Polish predominantly on the basis of a language corpus gathered in the years 2000–2003 in the course of longitudinal studies in the form of parental diaries and audio-recordings. All the children were recorded regularly for a period of time from a few months to two years. Apart from that, in the case of two Polish children, Kinga and Michał, a parental diary was kept. Those involved in making recordings and keeping the diaries were the children’s parents, including myself, who were asked to record spontaneous conversations with their children at home. Here is the detailed information concerning the children studied and the procedure adopted for compiling the corpus. 1. Polish children: Kinga, Zuzia and Michał. a. Kinga – tape-recorded from the age of 1;6 to the age of 3;6. From 1;6 to 2;6 she was recorded every two weeks for thirty minutes, from 2;6 to 3;6 every month, also for thirty minutes. In addition, a parental diary was kept in which both her utterances and their context were written down. b. Zuzia – tape-recorded from the age of 1;8 to the age of 2;1 every two weeks for thirty minutes. c. Michał – tape-recorded from the age of 2;2 to the age of 2;6 every two weeks for twenty minutes. Apart from that, his mother also made occasional notes of his utterances which were incorrect from the point of view of the adult user of the language. 2. English-speaking children: Ella and Jefferson (American English) and Jack (British English). During the process of collecting the language sample two of the children studied, Ella and Jack, lived in an English-speaking environment, in the USA and Great Britain respectively, whereas Jefferson lived in Poland but had hardly any contact with the Polish-speaking community. All the parents of these three children were native speakers of English. a. Ella – tape-recorded from the age of 1;2 to the age of 1;8 for half an hour every month. From the age of 1;8 to the age of 2;9 she was digitally recorded for half an hour every month or every two weeks. All the recordings were made by her father, who, if necessary, provided the context of her utterances or explained what the girl wanted to say. Ella was the most linguistically advanced of the six children studied, as her linguistic abilities considerably surpassed her biological age. In comparison with 11
other children, she exhibited the fastest rate of linguistic development and the best working knowledge of grammatical structures. b. Jefferson – tape-recorded from the age of 2;6 to the age of 3;10 for twenty minutes every two months. c. Jack – tape-recorded three times: at the ages of 1;8, 1;10 and 2;0 for half an hour. For the purpose of my analysis of the child language carried out in the chapters to follow I labelled the children studied in the following way: K stands for Kinga, Z for Zuzia, M for Michał, E for Ella, J for Jefferson and Jk for Jack. As far as the organisation of this publication is concerned, it consists of three chapters. Chapter I is devoted entirely to an overview of first language acquisition theories, concentrating on those that have become landmarks for the first language acquisition research. My aim is to show that each new development originated as an antithesis of the theory preceding it. The way in which the views on the nature of the process of first language acquisition evolved is a direct reflection of an everlasting dichotomy between those who support the idea of the innateness of language and its detachment from cognitive skills, and those who are of the opinion that language is not innate and that it is inseparably connected with cognition. Thus, I give a short account of behaviourism, which is believed to have spurred considerable interest in the study of child language. Then I briefly comment on the Chomskian Universal Grammar theory which originated from criticism of behaviourist approaches on the grounds that they were too simplistic. Lastly, I concentrate on the most recent developments, namely cognitivist accounts of first language acquisition, refuting the premise concerning the essential role of creativity in the process of first language acquisition put forward by generativists. In chapter II I concentrate on the children’s earliest utterances, formed at the onset of a multi-word stage, from which numerous important sentence elements have been deleted, and which is referred to as telegraphic speech. I attempt to account for the nature of these omissions in the light of both traditional approaches and the latest theories of first language acquisition. At the same time I come up with a wide variety of interpretations of such incomplete utterances, resulting from the phenomenon of language redundancy. Next, I carry out an analysis of telegraphic utterances both from the English and Polish sample with a view to specifying the degree of their ambiguity as well as establishing differences and similarities between the Polish and the English sample, as far as the grammar of telegraphic utterances is concerned. Finally, chapter III is devoted to the morphological and syntactic development that children undergo, analysed within the cognitive framework. As 12
far as both the morphological and syntactic analysis of the language sample is concerned I rely on the semiotic theory advanced by Charles Peirce, and the triad of linguistic signs proposed by him, i.e. index, icon, and symbol. With regard to morphological processing I adhere to the theory of Natural Morphology with the aim of establishing the degree of iconicity that the child language is characterised by. The syntactic structures produced by Polish and English-speaking children are analysed in the light of three iconic rules: of quantity, sequentiality, and distance. As can be seen, this work has as its ultimate goal the presentation of the early stages of morphological and syntactic development in child speech in two typologically different languages, i.e. Polish and English. It aims to describe similarities in the developmental pathway in these languages, rooted in the uniformity and universality of cognitive processes going on in the child’s mind irrespective of his native language as well as language-specific differences, resulting from the different typological characteristics of the two languages in question.
13
CHAPTER I
General overview of first language acquisition theories Introduction The majority of issues raised nowadays in the field of first language acquisition are deeply rooted in earlier linguistic, pedagogical and sociological developments. However, the aim of this chapter is not to provide an in-depth study of earlier approaches but rather to present chronologically the theoretical foundations of today’s thinking as well as the origin of the long-lasting dichotomy between the supporters of the innateness hypothesis and those who claim that language emerges in the course of overall cognitive development.1 For the most part I will restrict my considerations to the post-war period, which in its early years saw the rapid development of psycholinguistics, and brought about a breakthrough in first language acquisition research. Numerous theories were put forward then which later on constituted and still constitute a frame of reference for other studies. I will begin with a brief discussion of the conceptions dominant in mainstream psychology and linguistics in the 1950s and 1960s, which, because of their impact, may be referred to as milestones in first language acquisition studies. I will also view in retrospect, where necessary, earlier approaches. Then I will go on to describe the far-reaching influence of the ‘Chomskyan revolution’ on the study of first language acquisition in the 1970s and I shall also see that it has produced long-term consequences. Next, I will devote my attention to semantic approaches that led to the emergence of cognitive theories of first language acquisition. My aim will be to show that cognitivism has evolved from criticism of Chomsky’s Universal Grammar. Moreover, it is deeply rooted in the development of the notions of prototype and category. In the last part of the present chapter I will describe the whole array of cognitive approaches to first 1
An exhaustive analysis of early first language acquisition theories can be found in, for example, Selinker (1992) or Dulay et al. (1982).
14
language acquisition and show their numerous applications in today’s first language acquisition research.
1.1. Milestones in first language acquisition research 1.1.1. Earliest records of first language acquisition To begin with, according to Campbell and Wales (1970), the earliest recorded study of first language acquisition was made by the German biologist Tiedemann (1787), who carried out a general analysis of child development. Other important early studies include Charles Darwin’s (1877) and Hippolyte Taine’s (1877). However, it was not until the German physiologist Preyer (1882) made detailed daily notes of the first three years of his son’s development that the study of child language gained its true founding father (Campbell and Wales (1970)). According to Ingram (1989), Preyer’s study falls within the period of the so called diary studies (1876–1926), where the method of data collection was the parental diary in which a linguist would record his child’s linguistic development. Nonetheless, the Preyer’s study is not limited to linguistic development alone; it includes many notes on, for example, motor development and musical awareness. The first published book to be devoted exclusively to the study of child language was W. Stern’s Die Kindersprache (1907). It is from this work that the notion of stages of language acquisition, applied so widely in psycholinguistics, derives. However, only after World War I, with the rising popularity of behaviourism, did the study of first language acquisition begin to develop. 1.1.2. Behaviourism Although behaviourism gained widespread popularity in the 1950s, the foundations for behaviourist thinking in this period were laid as early as in the 1930s. It was then that the attempt of Leonard Bloomfield, one of the founders of the school of structural linguistics, to account for the child’s acquisition of word meaning received considerable attention. Structural linguists believed that [...] language systems consist of a finite set of ‘patterns’ or ‘structures’ which act as models for the production of an infinite number of similarly constructed sentences (Howatt (1988: 14–15)). In this view, the linguistic analysis of a sentence pattern both in adult and child language consists in classifying lexical items into classes, applying the criterion of distributional properties of words. New utterances are formed by analogy and inductive generalisations about the position of words in a sentence (Lewandowska-Tomaszczyk (1988)). This theory gave rise to distributional approaches to first language acquisition. According to 15
distributional learning theories, children seek the semantic or syntactic properties of words on the basis of their distribution, that is the range of positions in which they occur in a sentence. These approaches vary from those which reject abstract syntax (Kuczaj (1982 a)), to others which allow for the possibility of its gradual emergence (Maratsos & Chalkley (1980)). Bloomfield (1933) has stated that a hypothetical language learner possesses virtually no innate linguistic knowledge apart from the ability to hear, distributionally classify words into word classes and form new sentences analogous to those heard before. This strategy adopted by the child leads eventually to habit formation because successful attempts at speech are reinforced and established as correct language forms, whereas imperfect attempts are lost. Learning understood as the formation of habits is central to the behaviourist view of language acquisition (Skinner (1957), Thorndike (1932), Watson (1924)), which sees human beings as being exposed to a wide variety of stimuli in their environment and equipped with nothing but imitation skills. Skinner (1957) has proposed that language can be acquired through a process called instrumental conditioning which consists in establishing a link between a particular (linguistic or extralinguistic) stimulus and a desired response (the correct verbalisation). In order to develop this bond it is necessary to reinforce every correct utterance until it becomes automatic and involuntary, in other words, until it is established as a habit. He has maintained that learning syntax consists in imitating parts of sentences heard before, which, positively reinforced, become sets of associations between words (Skinner (1957)). Grammar, from this point of view, is regarded as a set of classes of words based on these associations, occurring in a serial, predictable order. Unfortunately, when seen in this light, the role of the child is reduced to that of a passive receiver of stimuli coming from the environment, capable merely of reproducing a finite set of patterns and deprived of the possibility to shape his language behaviour in a conscious way. In spite of these early critical remarks, behaviourism continued with major attempts, for example, Mowrer (1960), trying to explain language in behaviourist terms. Moreover, Jenkins & Palermo (1964) elaborated on Skinner’s ideas on how syntax can be acquired without resorting to any innate language faculty. 1.1.3. Alternatives to behaviourism At the end of 1950s radical behaviourism witnessed declining popularity when some researchers began to claim that at best behaviourism can explain
16
only some phenomena and exclusively in the holophrastic stage,2 to which the principle of imitation and positive feedback might be applied (Ingram (1995)). Still harsher criticism was engendered when generative linguistics, emphasizing the rule-governed and creative nature of human language, was developed, initiated by Chomsky’s publication of Syntactic Structures in 1957 (Chomsky (1957)). Moreover, the behaviourist theory was fiercely criticised by supporters of the developmentalist approaches to first language acquisition, such as Piaget, according to whom language is the result of general symbolic growth. (Piaget (1970), Piaget and Inhelder (1966), Piaget (1952)). Behaviourism was criticised severely by Piaget on the grounds of its oversimplicity and due to the fact that it ignores completely the existence of congenital potential and disregards the creativity of human language. Instead, he has proposed the constructivist view of language acquisition, which means that children are endowed with the capacity to construct their language and that linguistic development is preceded by cognitive growth. This process proceeds in stages, i.e. particular phases in cognitive and linguistic advancement are attained by the child and earlier structures are built upon through interaction with the environment. Piaget’s approach stresses the importance of the period of prelinguistic development (from birth to the age of two), during which the foundations for speech production and perception are established. This period is referred to as sensorimotor because the child solves problems by means of its sensory systems and motor activity. This term also implies that the child derives understanding of the world solely from his action patterns associated with objects. For example, in stage 3 of the sensorimotor period, called secondary circular reactions, the appearance of antecedents to later classes of concepts can be observed. Piaget (1952:234) argues: […] the secondary schemata constitute the first outline of what will become classes or concepts of reflexive intelligence: perceiving an object as something to shake, rub, etc. These are the first foundations of the later classification of objects and their characteristic features, which will be responsible for establishing the concept of the parts of speech. Still another development taking place also in the prelinguistic period and essential for the later emergence of speech is object permanence. This notion is defined as the child’s awareness of the fact that objects continue to exist although they are no longer seen. It shows that the infant’s reliance on symbols becomes more and more advanced and gradually replaces his sensorimotor schemata. Only when the child is cognitively mature enough, is it capable of acquiring its native language. 2 Holophrastic speech consists of one-word utterances; that is, holophrases which function as a complex idea or a sentence. For example, the intended meaning of the holophrase milk may be ‘Give me some milk’ (Dale (1975)).
17
As can be seen, Piaget strongly emphasises the interrelation of language and cognitive processes. He insists on the priority of thought over language, the former warranting the acquisition of the latter. In Piaget’s view, the establishment of a system of symbolic representation is a prerequisite for the emergence of a sound system and it is the child, who, having inherited cognitive capacities, creates a mental representation of the world in its mind and consequently masters its native language. The role of Piaget’s constructivist view cannot be underestimated since it has been the model for several subsequent theories of language acquisition of cognitive orientation.
1.2. Chomsky’s Universal Grammar Yet another flood of criticism of behaviourist theories of language acquisition came from Chomsky’s theory of Universal Grammar. As has already been stated, in the 1960s and at the beginning of the 1970s it was Noam Chomsky and his model of first language acquisition that was considered to be capable of providing a powerful means of describing and explaining children’s early grammatical development. The major assumptions of this theory, stemming from the harsh criticism of behaviourism, were put forward in Chomsky’s review of Skinner’s Verbal Behaviour published in 1959. First and foremost, the behaviourist hypothesis that children learn exclusively due to their imitative skills was totally discredited by Chomsky on the grounds of the creativity of language. Such common examples as: I goed or foots show clearly that children not only do not copy language from their environment but also are aware of the existence of grammatical rules and put this knowledge into practice. Another argument against behaviourism is that children produce a multitude of sentences never heard before so they cannot be just merely reproducing strings of words. Secondly, taking into consideration both the degenerate language input (spoken language is full of false starts, slips of the tongue, etc.) and also the complexity and abstractness of linguistic rules, it is astonishing that children manage to master the grammar system relatively quickly and efficiently. This issue has been referred to as Plato’s problem (Chomsky (1987)) and it points to the fact that it is hardly conceivable that some structural properties of language can be acquired just on the basis of language samples from the children’s environment. In order to account for these phenomena Chomsky claims that children possess an innate language faculty leading them in their acquisition of language. He subscribes to the maturationist view, according to which grammatical 18
principles become available to the child at some genetically determined time. This assumption is the major point of divergence between Chomsky and Piaget because the constructivist view does not see linguistic development as the result of maturation but, as can be seen, accounts for all the changes in linguistic development by proposing that linguistic structure is built up step by step by young language learners. In the UG approach, when faced with the body of speech, children are genetically programmed to discover grammatical rules underlying language and guided by their linguistic knowledge as to what these rules should look like. Since Chomsky’s theory gave a strong stimulus to psycholinguistics, especially in the field of first language acquisition, now I am going to present its main assumptions. First of all, according to the Universal Grammar (UG) theory, all human beings are biologically endowed with a universal set of principles and parameters which determine the form of human language and at the same time make all languages similar to one another. The Government and Binding theory proposed by Chomsky (1981, 1986a, 1986b) argues that these two elements are essential to the core of human language. Principles are unvarying and apply to all natural languages, underlying their structure, whereas parameters are language specific and because of that they determine the differences between particular languages. Chomsky (1981:4) states that […] what we expect to find, then, is a highly structured theory of UG based on a number of fundamental principles that sharply restrict the class of attainable grammars and narrowly constrain their form, but with parameters that have to be fixed by experience. Obviously, this is not sufficient to explain the process of first language acquisition. In addition, Chomsky proposes the existence of a genetically transmitted language faculty, which guides children in the acquisition of their native language and which is biologically triggered according to a predetermined timetable. Another issue inseparably connected with the UG approach, is that it is not concerned with performance, i.e. about how language is used in real life and influenced by interfering factors, such as time constraints or restrictions on the brain’s information-processing facilities. Instead, it deals exclusively with the abstract mental representation of language which all human beings possess, which is referred to as competence. Therefore, let us review briefly competence factors, that is, principles concerned with how the child constructs grammatical rules and changes them over time. When assessing language acquisition data, it is immediately obvious that children overgeneralise morphological inflectional suffixes to irregular forms, saying words, such as: foots and breaked. Stern (1983) refers to these forms as 19
child etymologies. To account for the existence of these forms in the child language Dresher (1981) has proposed the generalisation principle. According to this principle, with the passage of time, more complex forms that require more marking and are exceptional, are overgeneralised as regular, unmarked ones and preferred over them since their formation doesn’t require any radical changes to the structure of words. At some point in development, the generalisation principle restricts the form of the child’s rules by excluding all irregularities. Since the acquisition of irregular forms in English by means of the generalisation principle cannot be explained, the lexical principle, stating that individual paradigmatic alternations are learnt as separate lexical items, has been proposed (Ingram (1985)). It shows why children initially do not produce overgeneralisations, such as foots (they either omit irregular plurals or use them correctly). This is due to the fact that at first they acquire the paradigmatic variants like cat-cats, dog-dogs as separate pairs of words, only later realising that there exists a plural morpheme -s. In the light of this principle, in the beginning children produce irregular plurals correctly, treating them as separate vocabulary entries; however, after they discover the existence of the rule of forming plurals and the plural marker, they attach it indiscriminately to all singular nouns. However, by passing through these two stages in the acquisition of irregular forms, the child could conclude that there are two plural forms: foots and feet. In order to explain why this does not happen Wexler & Culicover (1980) have proposed the uniqueness principle, stating that the child selects only this form which is used in the child’s linguistic environment. Finally, children also resort to the principle of structure dependency in forming their utterances (Cook (1988)). This principle asserts that knowledge is based on […] structural relationships in a sentence rather than on a sequence of items (Cook (1988: 2)). It means that the knowledge of language involves the knowledge of syntactic categories. Proponents of the UG theory have presented several arguments from the first language acquisition data to support their hypotheses. First of all, the main characteristics of L1 acquisition share the same properties not only within one language but also across languages.3 Thus, irrespective of the language to be learnt, all children go through similar stages (e.g. the acquisition of negation is very similar cross-linguistically). Moreover, child language is rule-governed and systematic, and very often rules created by children do not correspond to adult ones. Children are also extremely resistant to correction, which is yet another piece of evidence that they possess an internalised system of rules that they stand by. Therefore, the 3
Evidently, as will be seen in the subsequent part of this chapter, this view is not shared by the proponents of cognitivism.
20
universalists concluded that these features of first language acquisition, so regular and systematic, must be the result of the existence of a genetically determined language faculty, guaranteeing that this process proceeds along the same lines in all children. Otherwise, it would not be possible to perceive such numerous similarities across languages since there would be too many individual variations (Mitchell (1997)). Secondly, according to the UG approach, language does not seem to be linked in any clear way to intelligence. Steinberg (1993) remarks that [...] many very famous people, including Albert Einstein are reputed to have been slow to talk. In support of this view there is also a large body of evidence from children with cognitive deficits whose language develops normally. Bishop and Mogford (1993) and Bellugi et al. (1993) have studied children suffering from Williams syndrome, a rare metabolic disorder which causes, among many other things, severe mental retardation. They clearly show that language development and cognitive prerequisites which, as argued by Piaget, are necessary for language acquisition are not mutually related at all. They came across a sophisticated use of language with complex syntax and advanced vocabulary in individuals whose mental development was below the level of a seven-year-old child. On the other hand, there also exists evidence of the opposite: that is, there are children who, although cognitively ‘normal’, suffer from sometimes severe language impairment. This condition, referred to as SLI (Specific Language Impairment), is characterised by language being deficient in some areas. Recently, an English-speaking family was studied and the results of this research conclusively prove that 16 out of 30 of its members in the last three generations exhibited SLI. Thus, it can be argued that it is an inherited disorder and therefore it can be stated that at least some aspects of language may be genetically controlled ((Cook (1997), Gopnik and Crago (1991), Pinker (1994)). For more detailed accounts of the relationship between brain and language see, for example: Caplan (1987, 1992), Harris and Coltheart (1986) and Sabourard (1995). To sum up, in the light of this evidence, universalists concluded that there must be some kind of innate language faculty which is biologically triggered, explaining why language simply matures in children in the same way as children naturally begin walking. Chomsky (1975), for example, saw no reason why intellectual development should be separated from physical. In his view, if the physical structure of the organism is biologically determined and taken for granted with such dimensions as size and rate of development also being inherited, why should we take a different approach to mental development? In fact, this kind of thinking was initiated by Lenneberg (1967) in his influential book entitled Biological Foundations of Language in which he 21
enumerated the characteristic features of biologically prompted behaviour and maintained that language complies with the rules governing physical development. Aitchison (1989:69) presents Lenneberg’s characteristics of language understood as an inherent part of biological development. The major theses of this argumentation are the following: 1. Language [...] emerges long before it is necessary. Children do not need to speak a language in order to survive because they are still being looked after. 2. Its appearance is not the result of a conscious decision. Similarly, children don’t decide consciously to start walking or sitting up. These are biologically determined processes which simply get activated in due time. 3. Its emergence is not triggered by external events (though the surrounding environment must be sufficiently ‘rich’ for it to develop adequately). Children need to be immersed in language to acquire it but on the other hand, there doesn’t exist any single event which can be regarded as a direct stimulus for children to start talking. 4. There is a regular sequence of ’milestones’ as the behaviour develops and these can usually be correlated with age and other aspects of development. It means that at first children begin to babble, then they go through a one-word stage, next they produce their first word combinations, etc., in the same way as children all over the world first try to sit up, then stand up, and then walk before they begin to run. 5. Direct teaching and intensive practice have relatively little effect. It is wellknown that children are resistant to correction (Mitchell (1998)) and although often provided with correct language input, they stick to their incorrect forms. 6. There may be a critical period for the acquisition of behaviour. This means that human beings need to be exposed to language before puberty if the language is to develop at all, in the same way as certain species of birds need to have some contact with their species’ song repertoire before a certain age so as to be able to produce it. The reasoning of the universalists presented above has allowed them to propose the existence of a language specific module in the brain in the form of a Universal Grammar consisting of a set of principles which apply to all languages and parameters which can vary from one language to another but only to a certain extent. The above presented line of thinking gave rise to the so called maturational theory of first language acquisition (Borer and Wexler (1987)). According to it, principles of Universal Grammar are genetically pre-programmed to become operational at different but strictly determined stages, just like any aspects of human development. For example, Wexler (1994) has stated that the notion of 22
grammatical tense matures at the age of 2;6. However, an apparent weakness of this theory is that the process of maturation cannot be empirically verified, and its validity therefore remains questionable.
1.3. First grammatical analyses The idea, proposed by Chomsky, that language is rule-governed gave rise to numerous theories of first language acquisition which proposed that the acquisition of language consists in the acquisition of grammatical rules of this language. These first grammatical analyses sought to explore the way in which children master the grammar of their native language. In what follows I will be concerned with the first ever grammatical analyses of children’s speech carried out during the period of first word combinations since it has been regarded by the majority of researchers as crucial for the onset of understanding grammar and, consequently, producing syntactic structures (Shipley, Smith & Gleitman (1969), Katz, Baker & Macnamara, (1974), de Villiers & de Villiers (1973a), Rodd & Braine (1970)).4 To define this period, let us have a look at the most frequently cited method of dividing grammatical development into stages as proposed by Brown (1973). Brown decided to use the average length of sentences as a means of differentiating between particular stages in the developmental continuum. In order to specify this, he decided to count the number of morphemes in children’s utterances because he believed that morphemes are the linguistic units which store more information about particular stages of grammatical development than just the number of words in the sentence or the age of children.5 This measure of the linguistic complexity of children’s utterances is commonly known as the child’s Mean Length of Utterance. Using the MLU as the criterion of dividing early linguistic development into stages, the onset of syntactic comprehension and the first attempts at production of word combinations take place when the average MLU is between 1.50 and 1.99, which is referred to as late Stage I.6 4
The experiments conducted by these researchers have proved that during this stage syntactic understanding takes place. For example, the data from de Villiers & de Villiers indicate that children are receptively aware of the role of English word order by late Stage I and also that they are capable of distinguishing between passives and actives. 5 Brown’s work was pioneering because of the fact that previously all other researchers measured the number of words in order to determine the stage of linguistic development. For example, Nice (1925) introduced the measure of the Average Length of Sentence (ALS) which was the mean number of words used by the child in spontaneous language samples. 6 According to Brown (1973), Stage I consists of early Stage I (1;0-1;49) and late Stage I (1;50-1;99) and is characterised by the beginning of the acquisition of the basic semantic relations
23
The first attempt to explain the child’s early grammar at Stage I, which received a great deal of attention, was that of Braine (1963a). He put forward the claim that early grammatical development can be accounted for within the framework of the theory that he proposed, that is pivot grammar. Pivot grammar puts the main emphasis on the importance of distributional learning and holds that the child initially selects a small group of words for grammatical acquisition on the basis of their frequency in the language input. These words occupy a fairly fixed position in a sentence. Unfortunately, one of the shortcomings of this theory is that it does not explain how the child selects these words. Nevertheless, this small class of words picked out by the child is known as the pivot class, whereas all the other words belong to the open class. Since there is a lot of individual variation among children acquiring their native language, each pivot class will also be different. The pivot class is an indispensable tool for acquiring open classes (this small group of words gets attached to other words). What is more, because the pivot class occurs only in two-word combinations, there are two possible pivots: sentence-initial pivots or sentence-final pivots, whose position is always restricted. The child is able to recognise that certain open words combine only with certain pivots. Pivots, however, may never occur alone or with other pivots, as this would contradict the assumption that they are used to acquire other words, that is open classes. The defining features of a pivot grammar have been presented in the table below: Feature
Pivot class
Open class
size frequency position co-occurrence isolation
small class frequent restricted cannot co-occur do not occur alone
large class infrequent not restricted may co-occur may occur alone
Thus, examples of pivotal constructions would be: all broke, all clean or no bed, no home, or I see, I shut but also see baby and see train (Braine 1963 a: Table 2). On the basis of these utterances it is evident that this hypothesis is internally inconsistent because, for instance, from the above data it is evident that I see is a combination of two pivot words, which is in violation of the theory. Another criticism of Braine’s grammar is that it lacks precision. Braine does not provide any procedures for determining the frequency of occurrence of particular classes; nor does he specify how they get established. used in language, such as Agent or Patient. Moreover, the child then becomes aware of the function of word order in sentences. Stage I is preceded by the period of single-word utterances (1;0) when single words are used without any grammatical knowledge.
24
Initially, pivot grammar gained wide appeal mainly because of its simplicity and the fact that at first glance it seemed to have succeeded in explaining the child’s early utterances. However, with the passage of time, it drew more and more criticism with Bloom (1971), Bowerman (1973 a) and Brown (1973) challenging its validity until it was discarded completely.7 Another study inspired by the discovery of the rule-governed nature of human language was Roger Brown’s so-called morpheme study (Brown (1973)). Analysing the speech of three children of different backgrounds, he compared the development of fourteen grammatical morphemes in English and subsequently found out that the order in which they were learnt remained constant in spite of the varying rate of their acquisition. Brown took as a criterion for acquisition the presence of a morpheme in 90% of its obligatory contexts in three successive 2-hour-long speech samples from a child. This order has been presented below (Brown (1973)): Present Progressive Prepositions: in and on Plural Past irregular Possessive Uncontractible copula Articles Past regular Third person singular Third person irregular Uncontractible auxiliary Contractible copula Contractible auxiliary
boy singing flower on window, dolly in car Sweeties Broke baby’s biscuit he is my friend a car Wanted Eats he has he is running he’s my friend he’s running
Remarkably, children not only acquire these morphemes in a fixed order but also within a given area of grammar they follow strikingly similar stages. Another interesting remark is that these stages are reached in a fairly invariable order across languages. For example, children all over the world not only acquire negatives at around the same age, but they also take similar steps in this 7
For example, Bloom (1970) pointed out the existence of ambiguous sentences in early speech, such as mommy sock, which occur in different contexts with different meanings and may mean ‘mommy is putting on her sock’ or ‘this is mommy’s sock’ or ‘mommy, I want a sock’, etc. Pivot grammar is not capable of accounting for these semantic differences. Further criticism of pivot grammar may be found in Blount (1969), Kernan (1969), Park (1970a, 1970b) and Rydin (1971).
25
process. Thus, at first they attach the negative marker to the outside of the sentence, producing the following utterances: no go to bed, or in French: pas faut boire (=not need drinking). The next stage is gradually to move the negative marker inside the sentence as in there no squirrels. Finally, the auxiliaries are acquired, which leads to grammatically correct utterances although some mistakes still occur, for example double negatives or omitting the copula be (see Ellis (1994: 78), based on Clima and Bellugi (1966) and Cazden (1972)). The fact that children possess an internalised system of rules was gaining more and more attention although it was made clear that the rules that children create may initially differ considerably from adult ones. As can be seen, even at the two-word stage children consistently express a variety of relationships between particular elements in a sentence. Moreover, they commonly come up with such forms as: milks or bringed, which they have never heard before so it is evident that they do not imitate the adult language but rely on their own grammar system. The evidence from Berko’s famous experiments, carried out with very young children in the 1950s, proves this conclusively (Berko (1958)). Berko developed a technique to study the acquisition of some grammatical morphemes in English, such as plural, past, etc., known as the wug procedure. First, she created a series of nonsense words and nonsense drawings which were supposed to represent meanings for nonsense words. Then the children were shown, for example, a picture of a strange, bird-like creature and were told: This is a wug. Next, they saw two such creatures in one picture with the experimenter trying to elicit the correct plural form: Now there is another one. There are two of them, there are two ________ . The children were expected to fill in the last word and the vast majority of them (91 per cent) replied: wugs, showing that they not only remember plural forms from previous utterances but that they also extract a plural rule from the language they hear and then apply that rule to their own productions. The experiment also checked the ability to use, for example, past tense morphemes by presenting to children a picture of a person performing some extraordinary action and telling them: This person knows how to gling. He is glinging. Yesterday he did the same thing. Yesterday he_____________ . Again most of the children responded correctly: glinged (77 per cent of them), in this way proving that they are familiar with the rule for forming the past tense. In the 1970s it was maintained that the outcome of these experiments again had provided evidence in favour of Chomsky’s innateness hypothesis and the Universal Grammar approach, proposing that the form of human language is genetically determined, which would explain why the process of first language acquisition is so uniform within one language as well as across languages. On 26
the whole, the UG theory was the dominant theory of first language acquisition in the 1960s and 1970s. It inspired many linguists who adopted its main assumptions in order to explain the process of first language acquisition. Even more recently, i.e. in the 1980s and 1990s, the considerable influence of the theory of UG can be observed. Therefore, the next subchapter will be devoted to the impact that UG had in the field of first language acquisition.
1.4. Theories stemming from the UG approach In the 1950s and early 1960s the only analyses of L1 acquisition being done were distributional, such as the above mentioned Braine’s pivot grammar. Obviously, they were not satisfactory from the linguistic point of view because their prime concern was rather the child data itself than its explanation or referring it to any theories of adult language. However, with Chomsky’s work becoming more and more influential, grammatical research in the period from 1965 to 1975 was dominated by the UG approach as well as the theory of syntax developed by Chomsky (1965), which has since been referred to as the Standard Theory, proposing the existence of a generative grammar whose main task is to describe all grammatical sentences of a language by a set of rules. The first person to use UG and Standard Theory as an explanation of language acquisition was McNeill ((1966a, 1966b), Miller and McNeill (1969), McNeill (1970a, b, 1971)). He claims that children are born with the blueprint for language in their minds in the form of a highly constrained universal grammar. The basic grammatical relations underlying all languages constitute the foundation of this grammar. According to McNeill (1970 b), these major grammatical relations are the following: Relation
Feature
Example
Predicate
[+VP, +NP_______ ]
[The dog ate the apple]
Subject
[ +NP, +______ VP ]
[The dog ate the apple]
main verb
[ +V, + _______ NP ]
The dog [ate the apple]
object
[ +NP, + V _______ ]
The dog [ate the apple]
modifier
[ + Det, + ______ N ]
[The dog] ate [the apple]
head
[ + N, +Det _______ ]
[The dog] ate [the apple]
27
The first feature stands for its grammatical category, whereas the second represents its subcategorization. For example, the relation main verb is represented by a V and it occurs before a NP which is its object. The brackets specify the domain of each relation, since all the constituents of this sentence also relate to the others. The task assigned to the child is, according to McNeill, to identify these relations in the language. They are not acquired simultaneously but in a fairly predictable order during the period of first word combinations. There is an obvious similarity between Mc Neill’s theory and Braine’s approach because McNeill accepts the superficial features of the early pivot grammar. According to him, the child’s initial grammar looks almost like the pivot grammar but, eventually, the child ends up with a basic phrase structure grammar with the universal grammatical relations. However, unlike in Braine’s theory, the initial pivot class is not an accidental collection of words determined on the basis of their frequency and positional constancy but is constrained by the hierarchical nature of the universal grammar relations. This means that the words are grouped at the higher level of organization. For instance, predicate, main verb and modifier occur with NP and therefore a pivot class is, in the beginning, a group of words which can be combined with nominals. The acquisition of grammatical relations begins as early as during the holophrastic period. McNeill (1970b:1096) states: It is important to note that these features are automatically made available when the child obtains any meaning from adult speech. What McNeill implies here is that children make use of semantics at this stage and that they know a range of semantic categories. Therefore, he argues that these semantic categories and their use in actual speech are constrained by the universal grammatical relations. This process leads directly to the establishment of syntactic categories because the child assigns the relevant grammatical feature to each new word and the Universal Grammar possesses a set of principles determining the phrasal structure of a sentence. In other words, any constituent that, for example, has the feature [+N] or [+V] is automatically part of an NP or VP, respectively. This discussion is also part of the theory known as X-bar syntax (Jackendoff (1977)). This kind of development leads to creating a grammar which is adult-like from the very beginning with regard to its representation of grammatical relations. McNeill (1970b:1096) argues: Insofar as the basic grammatical relations reflect the innate abilities of children, the type of grammar just outlined will be developed regardless of the language to which the child is exposed. It is a universal child grammar. While McNeill accounts for the emergence of syntax, some other attempts were purely descriptive in nature. The first to use Standard Theory as 28
a descriptive tool was Menyuk (1969) who wrote transformational grammars for a large sample of children, putting a lot of emphasis on errors and omissions that children make. However, these samples were too scanty for individual grammars to be presented. The first in-depth attempt to write individual grammars was that of Bloom (1970), in which she used the Standard Theory to describe the child’s data. As a result, she formulated a set of grammars for three subjects. Bloom’s grammars fall somewhere between those of Braine and McNeill; that is, the data from early stage two is similar to Braine’s theory, whereas that of late stage two resembles more McNeill’s grammar. Bloom’s prime concern was, however, to prove the existence of grammatical relations that McNeill described on the basis of the child language. This was the first attempt to verify McNeill’s theory empirically (detailed characteristics of Bloom’s theory can be found in Ingram (1989: 272)). When it comes to more recent theories rooted in Chomsky’s line of thinking, three approaches based on the issue of the relationship between child and adult grammar can be differentiated. These are: 1. The Strong continuity approach (Weissenborn (1990), Poeppel and Wexler (1993), Hyams (1992, 1994, 1996)). In this view the child’s grammatical system is practically identical to the adult system. The only difference between them results from the fact that some phonetic elements are not overtly realized. As Hyams (1994) has stated, children show evidence of syntactic relations at a stage when they are incapable of producing lexical elements, considered to be the carriers of syntactic information. The failure to produce lexical items can be traced back to the fact that they typically lack referentiality or meaning. 2. The Weak continuity approach. According to this, principles of UG are available to the child from the very beginning of the acquisition process, however, the child’s system may deviate form the adult’s in two ways. First of all, it may constitute a subset of the adult system (Clahsen (1990)). Secondly, it may be underspecified with respect to it (Lebeaux (1998), Radford (1990)). 3. The Discontinuity approach (Bickerton (1990)). In this view, at an early stage the child’s language is completely different from that of the adult’s. Bickerton (1990) refers to this initial linguistic system as ‘protolanguage’, claiming that it possesses no proper linguistic characteristics. He states that protolanguage is not structured in any way and that it merely consists of strings of words, similar to those produced by trained chimpanzees. In his opinion, the reason for this is that the principles of Universal Grammar have not emerged yet. However, UG theory and the theories stemming from it turned out to be unsatisfactory in explaining several aspects of first language acquisition, e.g. the 29
question of vast individual differences in linguistic development. Also, there is the lack of empirical evidence in favour of UG’s main assumptions. Therefore, new approaches, such as minimalism, and semantic and cognitive theories have developed. The remaining part of this chapter will be devoted to a discussion of these more recent developments.
1.5. New developments 1.5.1. Minimalism and functional categories As the UG approach was sharply criticized by its numerous opponents who claimed that it is too abstract and impossible to verify on an empirical basis, in the 1990s new theories aimed at the explanation of the first language acquisition process were put forward. The first theory I am going to discuss at this point is the Minimalist Programme. It was proposed by Chomsky (Chomsky (1995), Cook and Newson (1996), Radford (1997)) and it became quite influential in the 1990s because of its novel approach to the definition of parameters and their interrelation with principles. In this view the principles underlying language are fixed in advance, as before. However, parameters are neither related to specific principles nor are they part of the structural grammar, as was previously argued, but they are contained within the lexicon. As a result, languages differ greatly from one another only because of their miscellaneous lexicons and, consequently, language acquisition consists basically in learning the lexicon. This lexical parameterization hypothesis suggests that the parameters are contained within special categories in the lexicon, referred to as functional categories. Functional categories contrast clearly with lexical categories, that is the so called content words, such as nouns, adjectives, verbs, etc., which carry a specific meaning and belong rather to the lexicon than to the grammar. Function words are grammatical words, such as determiners, complementizers or grammatical morphemes. Another possible criterion which may be applied for the purpose of making this division would be assigning all lexical items either to closed or open classes of words. An open class of words stands for lexical categories and this means that new words can be and, are, in fact, added to it quite freely. For example, with the advent of new inventions and the development of technology, our vocabulary becomes enriched constantly with such words as: snowboard, e-mail, to computerize, etc. In contrast, a closed class of words, that is functional categories, consists of a fixed number of elements which are invariant and to which new words cannot be easily added. For
30
instance, it is impossible to add new determiners or new present tense morphemes to the language in the same unlimited way as new adjectives or nouns are added. In fact, this division into content and function words had already been recognised in linguistics for some time. However, what is new in this more recent approach is that function words or morphemes occur with some phrases attached to them in the same way as lexical words do. As a result, these are functional phrases which are organized like any other phrase, with the function word or morpheme as the head of this phrase. Consequently, there are, for example, Determiner Phrases (DP), Complementizer Phrases (CP), Tense Phrases (TP), etc. In acquisitional terms, the minimalist approach means that all the information concerning the language and its internal structure is stored in the lexicon. In fact, a lot of arguments in favour of this approach come from L1 acquisition studies. In this model, children acquiring their native language have to learn its lexicon, that is, both its functional and lexical categories. According to Radford (1990), children go through a stage of acquiring the lexical categories alone. L1 acquisition studies show that at around the two-word stage, which is sometimes also referred to as telegraphic, children’s knowledge of language manifests itself in knowing content words and their language seems to be devoid of functional categories, such as tense markings, determiners, etc. For that reason this stage is also sometimes termed as pre-grammatical with such utterances as: play doll, baby drink, daddy garden, etc. The explanation of this phenomenon, as understood by the minimalist framework, is that functional categories controlling a lot of the surface grammar have not been acquired yet. Although the minimalist approach can definitely be considered less abstract than Chomsky’s earlier developments, its theoretical framework still turned out to be unsatisfactory as far as several aspects of first language acquisition are concerned. First of all, the opposition between competence and performance has been sharply criticised. Chomsky proposes that children possess adult language competence, and the performance factors cause the production of utterances that are far from adult-like. However, the point is that this assumption has never been verified empirically and therefore it cannot be taken seriously. Secondly, virtually no one, apart from Pinker8, has put forward a theory aimed at accounting for the transition from the knowledge of Universal Grammar to the elements of the acquired language. Thirdly, generativists strive to explain exclusively grammatical 8
Pinker (1984) has formulated the Lexical Learning Approach, according to which grammatical acquisition is propelled by the learning of lexical items with their specifications, such as, for example, mass noun or intransitive verb. For example, the lexical entry for give will specify three arguments, that is agent, theme and goal, realized as subject, direct object and indirect object.
31
phenomena belonging to the so called core grammar, neglecting all less canonical structures, i.e. those belonging to the periphery of language. As a result, in response to the disappointment caused by unsatisfactory explanations of linguistic phenomena offered by Chomskyan approaches, a new paradigm of theoretical linguistics has developed in recent years. It is referred to as cognitive-functional linguistics, and it seeks to specify the nature of human competence from a less mathematical and more psychological point of view (Bybee (1985), Croft (1991), Fillmore (1985), Givon (1985), Goldberg (1995), Lakoff (1997), Langacker (1987), Talmy (1998), Van Valin (1991)). According to this new approach, human competence comes down to mastering manifold linguistic symbols and schemas, constituting a linguistic form (signans) related to some communicative function (signatum). Cognitive-functional theories attempt to account for all the aspects of linguistic competence, from highly canonical (core) to highly peculiar (periphery) (Kay and Fillmore (1999)). Generally speaking, cognitivitism has become the prevailing trend in the field of first language acquisition nowadays. Therefore, the remaining part of this chapter will be devoted to presenting its main assumptions and theories, as far as child language development is concerned. 1.5.2. Cognitive theories of first language acquisition 1.5.2.1. Introduction
To begin my review of cognitive theories of first language acquisition, let us recall once again a distinction that has already been discussed, between linguistic competence and performance. As has already been stated in the preceding subchapter, UG linguists’ primary interest is in competence, that is in the construction and form of grammar underlying all linguistic production. This model is not concerned with performance, that is the way in which this linguistic knowledge is acquired and put to use. In contrast, cognitivists are more interested in performance, although they do not neglect the role of linguistic competence, regarding both competence and performance as closely related and interacting aspects in the process of language acquisition. In fact, the cognitive approach to language acquisition can best be summarised by the following statement: Human intelligence comes from both having the right knowledge and making it available at the right time (Anderson (1993:69)). However, the main difference between cognitive and linguistic approaches to language acquisition lies in the fact that cognitivists view language as inseparable from other aspects of cognition and what is more, they claim that a certain degree of cognitive maturity is indispensable for the process of language acquisition to take place. In contrast, UG theorists believe that human beings are 32
endowed with an inborn knowledge of the grammatical structure of languages, which becomes accessible as a result of the exposure to the language input and which unfolds irrespective of cognitive mechanisms. In this section I am going to focus on general cognitive development seen as the prerequisite for the emergence of semantic relations as well as analyse the most influential cognitive theories of first language acquisition. To begin with, cognitive linguistics developed in the late 1970s. It sees language as part of the cognitive abilities of the human mind, such as perception, memory, attention, reasoning, etc. (Malmkjær (2004)). In contrast with the traditional approach to language rooted in the Aristotelian classical definitions of categories, cognitive linguistics adopts a phenomenological approach as its philosophical basis (Lakoff and Johnson (1980, 1999)). According to this approach, the individuals’ access to the world or their consciousness is realized by their bodily experiences of that world (Geeraerts (1985)). Secondly, cognitive linguistics opposes Saussurean structuralist axioms, especially dichotomies, such as syntax vs. semantics or lexis vs. grammar, etc. Moreover, the principle of the arbitrariness of the linguistic sign has been replaced by the principle of motivation and iconic principles of linguistic organization. Thirdly, cognitive linguistics is opposed to generative linguistics, as it does not see language as an autonomous system detached from any other type of knowledge. In contrast, it holds the view that there is no definite distinction between linguistic knowledge and any other type of knowledge (Haiman, (1980)). As Goldberg (1985: 5) has put it: Knowledge of language is knowledge. As far as first language acquisition is concerned, the main body of cognitive-oriented research did not begin until the 1980s. However, as early as in the 1970s some linguists tried to adopt a cognitive approach to the phenomenon of acquiring language. This found its expression in proposing semantic theories of first language acquisition. 1.5.2.2. The origins of cognitivism: semantic approaches
First of all, several linguists dealing with semantic development have argued that the internalisation of sensorimotor schemas into mental representations, which usually takes place at the sixth stage of sensorimotor development, is a necessary prerequisite for the emergence of multiword utterances (Ingram (1974, 1978) Moerk (1975) Morehead & Morehead (1974), Sinclair-de-Zwart (1971)). The cognitive achievements postulated to occur at this stage include: the concept of object permanence, realising that whole categories of objects can be used for the same action and that different actions can performed by means of one and the same object as well as understanding the causal relationship between objects whose result is initiating actions and making changes in the objects that are acted upon. 33
As far as the order of acquisition of general semantic categories is concerned, it has been pointed out that semantic notions do not necessarily emerge in the child language in the same order as they are acquired non-linguistically (Fodor, Beve & Garret (1974)). Nonetheless, according to Leonard (1976) it is possible to present a plausible account of the order of acquisition of general semantic categories based on the level of their cognitive difficulty. Thus, the earliest semantic notions to appear are the operations of reference, such as, notice, nomination or recurrence requiring nothing but the understanding of objects which is achieved in the third and fourth stages of the sensorimotor period (Brown (1973)). The next stage is the specification of agents and objects, which, however, requires an understanding of the relationship between objects and events as well as familiarity with the concept of causality, i.e. that all people are potential sources and recipients of actions. Moreover, this stage is also characterised by acquiring semantic notions, such as attribution and possession, reflecting the child’s ability to represent relations between particular objects independent of relevant action. These cognitive achievements take place during stages 5 and 6 of sensorimotor development (Piaget (1954)). Lastly, notions of experience and experiencer emerge which refer not to overt activity but to a psychological state. Since these are abstract concepts, they emerge relatively late in Stage I (Leonard (1976)), as the child in the sensorimotor period is capable only of understanding and processing such information whose source is overt action. Another issue inseparably connected with the relationship between cognitive development and semantic relations is the question of mapping. Some theorists have argued that linguistic semantic categories map directly onto cognitive ones (Sinclair-de Zwart (1971)). In this approach, early meanings are considered to represent a set of internalised sensorimotor schemata rather than a set of linguistic deep structures (Bates (1976)). On the other hand, other linguists have suggested a sharper distinction between conceptual and linguistic knowledge. For example, Bloom (1973:121) does not believe that cognitive categories develop […] in a one-to-one correspondence with eventual linguistic categories. According to Bloom et al. (1975) the cognitive representations of the relationship among objects and events are more global, not distinguishing between the respective semantic roles of the elements, but rather representing the entire relationship. She claims (cf. Bloom et al. (1975:30)) that […] such differential semantic categories as agent, place, affected object, etc. are linguistic inductions that the child has made on the basis of his linguistic experience. Similarly, Bowerman (1975) suggests that the child’s general understanding of object permanence, causality or actions as distinct from the objects affected tells the child hardly anything about how to form semantic categories out of the words he is exposed to. 34
However, Schlesinger (1974, 1982) rejects this theory, proposing that language encodes various events and relationships and that it interacts with the child’s non-linguistic experience in classifying words on the basis of their belonging to particular semantic categories. He states that: [...] by hearing sentences in which all agents are treated the same way, the child acquires the agent concept with rules for realising it in its speech (1974:45). The study conducted by Golinkoff and Kerr (1978) which tested the ability of presyntactic infants to discriminate between objects performing actions and objects that have been acted upon seems to show that this is the case. The children (aged from 15 to 18 months) taking part in this experiment were presented with a film showing two actors: agent A pushing recipient B. After some time the roles were reversed: B became the agent and A became the recipient and this role reversal was perceived by the majority of children. Nevertheless, some other experiments have made it clear that cognitive factors alone are not sufficient to account for the child’s acquisition of word meaning. This has been demonstrated in a series of works by Johnston and Slobin (1979), Johnston (1985). Johnston and Slobin (1979) examined the acquisition of locative expressions expressing the following concepts: ‘in’, ‘on’, ‘under’, ‘beside’, ‘between’, ‘back’ and ‘front’ in four languages: English, Italian, Serbo-Croatian and Turkish. On the basis of Piagetian research (Piaget and Inhelder (1967)) they predicted the following order of acquisition regardless of linguistic community provided that cognitive development were the sole determinant of acquisition: Predicted order of acquisition and justification (Johnston and Slobin (1979)): 1. ‘in’, ‘on’, ‘under’: these are the first spatial concepts of containment, support and occlusion. 2. ‘beside’: a purely spatial proximity relation, not dependent on the speaker’s viewpoint 3. ‘frontf’, ‘backf’ (i.e. of objects that have inherent fronts and backs, e.g. houses): proximity to an inherent feature. 4. ‘between’: coordination of two proximity relations. 5. ‘front’, ‘back’ (in relation to objects without inherent fronts and backs, e.g. balls): coordination of the relative proximities of the speaker, reference object and located object. Forty eight children aged from 2 to 4 were tested in each of the four linguistic communities to check the validity of the above assumptions. The results of the experiment apparently seemed to confirm the cognitive predictions because four of the words, ‘in’, ‘on’, ‘under’ and ‘beside,’ were always acquired 35
before the others and the acquisition of ‘between’, ‘backf’ and ‘frontf’ preceded the acquisition of ‘back’ and ‘front’. However, some variability within the groupings was observed, e.g. the English children acquired ‘between’ before ‘backf’, whereas the Turkish children showed the opposite order. In order to account for the cases where sheer cognitive complexity is not enough to account for the order of acquisition, Johnston and Slobin have proposed that another factor intervenes there, that is linguistic complexity. The way in which linguistic complexity may affect first language acquisition will be presented and described in detail below. 1.5.2.3. The perceptual saliency approach
The study of linguistic complexity and its influence on language acquisition resulted in the putting forward of the perceptual saliency approach, which was the outcome of Dan Slobin’s research carried out in the 1970s and 1980s that ended in the publication of a two-volume cross-linguistic study of child language development in 1985. As far as the notion of cognitive salience is concerned, it can be traced back to the theory of apperception put forward by Rozwadowski (1903), according to which during the process of perception one concentrates on the dominant feature of the linguistic unit.9 Slobin accounts for the similarity in linguistic development across children and across languages by claiming that human beings are programmed to perceive and organize information in certain ways. What makes language acquisition possible in this model is perceptual saliency, which means that [...] certain linguistic forms are more ‘accessible’ or more ‘salient’ to the child than others (Slobin (1979:107)). On the basis of this assumption, in the 1979 edition of his book Psycholinguistics he formulated five operating principles which guide children in their processing of the linguistic strings that they encounter. He also listed five resulting universals, completely different from linguistic universals because of the fact that they are cognitive in nature and enable us to understand the way in which children perceive and organize their linguistic environment. The five operating principles (Slobin (1979:108–9)): Operating principle A: Pay attention to the ends of words. Operating principle B: There are linguistic elements which encode relations between words. Operating principle C: Avoid exceptions. Operating principle D: Underlying semantic relations should be marked overtly and clearly. Operating principle E: The use of grammatical markers should make semantic sense. 9
36
Rozwadowski’s theory is discussed in detail in chapter III.
Language acquisition universals based on these principles are as follows: Universal 1 (based on principles A and B): for any semantic notion grammatical realizations as postposed forms will be acquired earlier than preposed forms. In practical terms, this means that the acquisition of suffixes is easier than the acquisition of prefixed morphemes. Universal 2 (based on C): the following stages of linguistic marking of a semantic notion are typically observed: (1) no marking, (2) appropriate marking in limited cases, (3) overgeneralization of marking, (4) full adult system. This set of procedures is in fact applied to the acquisition of irregular verbs in English and is referred to as the U-shaped curve of learning (for further explanation of this phenomenon as well as some examples, see page 42). Universal 3 (based on D): the closer the grammatical system adheres to one-to-one mapping between semantic elements and surface elements, the earlier it will be acquired. This would suggest that, for example, regular plural noun forms are acquired earlier than arbitrary irregular ones. Universal 4 (based on E): when selection of an appropriate inflection among a group of inflections performing the same semantic function is determined by arbitrary formal criteria (e.g. phonological shape of word stem, number of syllables in a stem, arbitrary gender), the child initially tends to use a single form in all environments. For instance, Polish children tend to use the feminine gender of adjectives with masculine nouns ending in -a which is a typical ending for feminine nouns and they produce the following forms: duża mężczyzna ‘big man (fem.)’, zmęczona kolega, ‘tired colleague (fem.)’ etc. Universal 5: semantically consistent grammatical rules are acquired early and without significant error. By 1985, this list of operating principles had been enlarged, drawing on data from L1 acquisition in several languages, until it reached forty. However, the above early principles are sufficient to provide us with an outlook on first language acquisition within the framework of the perceptual saliency approach because they are fundamental to the whole theory and include its main assumptions. 1.5.3. Modern cognitive approaches to first language acquisition 1.5.3.1. Prototype and category
Apart from perceptual salience, which is of considerable importance for cognitivists, other issues central to cognitive theories of first language acquisition are the notion of prototype and the notion of category. The first to embark on the cognitive-oriented study of categories was E. Rosch, who initiated the beginnings of the interest in linguistic ontogenesis. 37
In 1977 she carried out a study devoted to the issue of the so-called basic level categories (Rosch (1997)). One year earlier Rosch and Mervis (Rosch et. al. (1976)) observed that categories which are both linguistically and cognitively more salient than others are the first to be acquired by children under three. The experiment conducted by Rosch has inspired several first language acquisition researchers disappointed with numerous versions of generative grammar put forward by Chomsky, considered to be too theoretical and out of touch with reality. Therefore, in the 1980s several linguists turned to the field of cognitive linguistics inspired by the work of its founders, that is, Langacker (1987) and Lakoff (1980). The majority of the early research that laid the foundations of cognitive linguistics concerned lexical semantics. Therefore, the earliest cognitively oriented studies of language acquisition also concentrated on the acquisition of lexicon by children (Dąbrowska and Kubiński (2003)). Consequently, the issue of lexicon, central to all the branches of cognitive linguistics, has inspired the emergence of two notions: the notion of prototype and the notion of radial category. The notion of prototype was formulated by Eleanor Rosch (1978) and contrasts with the classical Aristotelian definitions of categories assuming that all members of the category share some essential features, i.e. that all category members have an equal status and that category boundaries are clear-cut. Rosch has proposed a prototype view of categorization on the grounds of the fuzziness within or between categories. She claims that categories do not reflect objective features but are, instead, approximations consisting of clear, central, or, in other words, prototypical, i.e. most representative members. As members of a given category are not characterized by equivalent status, category boundaries are not clear-cut. Statements about central members of the category are processed much faster than statements about marginal members. Moreover, reasoning about any category is based on what is considered to be a typical example of this category. Numerous researchers have tried to rely on the notion of prototype to study the early stages of first language acquisition. Barrett (1995), Meints, Plunkett and Harris (1999) have discovered that initially children name only prototypical items belonging to a given category. Consequently, the early lexical mistakes are underextensions, with overextensions appearing only in the later stages of linguistic development. Moreover, the presence of prototypical elements has also been observed in grammatical ontogenesis (Taylor (1989), Marchman et al. (1991), Meints et al. (1999)) although no systematic research has been done in this field so far. As far the notion of radial category is concerned, it is connected to some extent with the notion of prototypicality. Lakoff (1987) has resorted to it in order 38
to analyse the internal complexity of lexical items. However, some first language acquisition researchers are of the opinion that the acquisition of grammar may also take place by applying the radial strategy. For example, Sinha et al. (1994) have carried out a comparative study of the acquisition of locative particles in English and Japanese and found that children make use of a strategy that can be referred to as radial. This means that they first acquire the least semantically and cognitively complex locative particles only to master more semantically and cognitively complex elements in more advanced stages of their linguistic development. Another issue inseparably connected with the development of lexical semantics in the 1980s and 1990s concerns the comparison of lexical and grammatical development. Attempts to explain lexical development within the generative framework have shown that it is incapable of offering a plausible theory of vocabulary acquisition. Adopting a generative theory of first language acquisition it would have to be assumed that lexical categories are innate and are accompanied by various options, i.e. parameters for various languages (Lightfoot, (1989), Crain and Wexler (1999)). Evidently, such a standpoint when applied to lexical development, seems absurd as it would imply that all lexical items, e.g. coca-cola, scanner, or immobiliser are innate. Since this seems hardly plausible (although some extreme nativist approaches, such as Piatelli-Palmarini (1989), Chomsky (1991) suggest such a possibility), it must be assumed that children’s cognitive abilities allow them to create complex semantic structures. If so, that is if children are able to acquire semantic structures, it seems quite likely that they are able to acquire syntactic rules as well. Consequently, cognitive linguistics proposes that the syntax of the language is not innate but is actively constructed by the child, discarding in this way the main assumption of the generative tradition. Cognitivists have invalidated the theory of generative grammar on many grounds. First of all, they have refuted the existence of linguistic universals and criticised the fact that generativists minimized the significance of features characteristic of individual languages. In the generative tradition linguistic universals constitute the core of grammatical description, whereas differences between individual languages are dismissed to the grammatical periphery. The core-periphery opposition has become a permanent feature of Chomsky’s numerous models of generative grammar. Having analysed the differences between particular languages, first language acquisition researchers adhering to the cognitivist tradition have stated that differences between particular languages are fundamental, while syntactic categories and relations are not universal (Croft (2001), Tomasello (1995)). Tomasello (1995) claims that the existence of linguistic universals is acknowledged only by those linguists who have adopted 39
a generative theory of language. Dąbrowska (2000a) has stated that the majority of languages manifest features characteristic only of these languages, not any others, e.g. the semantic functions of particular cases are characteristic only of Polish, not of any other Slavic language. Yet another argument against generative theory rests on the nature of first language acquisition. Generativists have claimed that the process of acquiring the language is uniform, which means that children come through exactly the same stages and acquire particular linguistic structures in the same order, while individual differences are so slight that they can be neglected. However, empirical data has invalidated this theory, as in the course of numerous studies it has turned out that individual differences are tremendous (Nelson (1981), Goldfield and Snow (1989), Richards (1990), Bates, Dale and Thal (1995)). These differences have been found in every field as well as at every stage of language acquisition. Wells (1985) has discovered that the differences in ‘the linguistic age’ can amount to as much as 30–36 months. Moreover, there is a lot of variation in the way in which children acquire their language. Analytical children master single words, mainly nouns, that they later learn to combine with other words. Holistic children acquire the whole phrases which are then broken up into single words and morphemes (Nelson (1981), Peters (1977)). Some children exhibit the knowledge of inflections as early as during the holophrastic stage, while others form sentences, using uninflected words. Besides, some children acquire basic morphological and syntactic rules at the same time (Smoczyńska (1985)). 1.5.3.2. The usage-based model
Another reason for the critical attitude to the generative theory of first language acquisition is its standpoint on linguistic creativity. Cognitivists are of the opinion that the role of linguistic creativity has been overestimated by generative approaches (Lieven et al. (1997), Tomasello (1992), Dąbrowska (2000a)). For example Lieven et al. (1997) regularly recorded one child for five hours per week for six weeks with a view to establishing the origin of multiword utterances in the child’s speech. It turned out that 63% of the utterances were repetitions of adult utterances just heard by the child and 27% could be derived from syntactic operations based on the usage-based model of language acquisition put forward by R. Langacker, such as gap filling: Where’s X = Where’s the butter? or the addition of the element: I got one + here = I got one here. Just 7% of all the multiword utterances consisted in performing two operations, e.g. gap-filling combined with the addition of an element, and only
40
2% of the utterances required three or more operations. Consequently, it can be stated that children acquire prefabricated fragments of language and their utterances assume the form of highly stereotypical formulas. As far as linguistic creativity is concerned, it emerges little by little, and various parts of the grammatical system often develop asynchronically (Richards (1990), Gathercole et al. (1999), Tomasello (2003), Dąbrowska (2000b)). At this point let us present the cognitive models of language acquisition formulated to replace the generative approach. As has been remarked above, R. Langacker (1987, 1988, 2000) has proposed the existence of the usage-based model of language acquisition. In this model language is considered to consist of conventional linguistic units of various sizes, degrees of abstraction and entrenchment.10 These linguistic units are abstracted from usage events, i.e. the cases of their being used in the communication process. As the child acquires more and more linguistic units of a particular type, i.e. formulas, they are converted into symbolic construction schemas by the reinforcement of their recurring common features. In this process the key role is played by the frequency of occurrence. Individual differences result partly from the differences in the language input aimed at children and partly from the size of the acquired units, which is probably a consequence of the differences in the phonological memory. Another concept that cognitive theories of language acquisition concentrate on is the concept of construction (Langacker (2000)). In cognitive grammar the construction is defined as the combination of the specification of form and the specification of meaning (Taylor (2001: 271)).11 However, it is construction grammar that has elevated construction to the central position in language (Fillmore (1988), Goldberg (1995), Croft (2001)). Since, according to Langacker (2001), there are no fundamental contradictions between construction grammar and cognitive grammar it is possible to adopt its definitions and assumptions for the discussion of the cognitive approach to the first language acquisition.12 Croft (2001) has stated that grammatical categories and roles do not have a universal character. They assume different forms in different languages. What is more, there exist different applications of the same form in one language, e.g. the transitive subject differs from the intransitive subject. Therefore, not grammatical categories and relations but constructions are the basic units of the language. By the same token, constructions are also basic units subject to 10
According to cognitive linguistics, by linguistic units I can understand morphemes, lexemes, or even whole phrases (Langacker (2000)). 11 Translation mine. 12 A detailed comparison of cognitive grammar and construction grammar may be found in Langacker (2001: 1).
41
acquisition (Tomasello (1998), (2000), Dąbrowska (2000b), Goldberg (1999)). This assumption has been verified empirically, as it has been found out that the earliest children’s utterances are made of single words, or indivisible and invariable phrases, such as: nie+ma ‘there+isn’t’, co+to? ‘what’s+this’, nie+chce ‘doesn’t+want’ (Dąbrowska and Kubiński (2003)). From the eighteenth month onwards children produce two-, three-, and later on, multiword utterances. According to Tomasello (2003), it is not the case that children first learn single words and only then combine them to create phrases or sentences. Tomasello claims that the vast majority of all the utterances heard by children, i.e. about 90%, consist of more than one word. All these utterances are placed in a specific context; therefore, children learn the combination of an utterance with a communicational intention. Another point is that children abstract smaller units, such as words or lexemes from larger units, such as phrases or sentences. These two processes play a principal role in the process of first language acquisition. In order to present the way in which this model operates let us analyse the process of acquiring the phrase do domu ‘home’ and combining it with other sentence elements (Dąbrowska (2003)). Having heard a few adult utterances containing the phonological fraction domu ‘home’, e.g. Idziemy do domu ‘we are going home’, Chcesz iść do domu? ‘Do you want to go home?’, the child may acquire this form and use to express the meaning ‘I want to go home’. With the passage of time children master more complex and partially schematic constructions, such as OSOBA do domu ‘PERSON home’, where OSOBA schematically specifies expressions that may occur in this construction. Consequently, children come up with the following utterances: ciocia do domu ‘aunt home’, pan do domu ‘man home’, Kasia do domu ‘Kate home’. What should be emphasised here is that the formation of utterances of this type is not based on grammatical rules (ciocia + do + dom + -u) but on the combination of the complex lexical unit do domu ‘home’ with a simple ANIMATE unit, e.g. ciocia ‘aunt’. Only in more advanced stages of cognitive and linguistic development do children break whole units up into their basic elements and, consequently, acquire more abstract schemas, e.g. do + NOUN + GENITIVE ENDING -u. Thus, in the light of what has been said above, abstract construction schemas emerge from fixed phrases or formulas acquired earlier by the child.13 Evidently, the following question arises: what makes the child pass from the formula to the schema? Dąbrowska (2003) has tried to explain the nature of this passage, carrying out an analysis of the acquisition of questions in 13
42
A detailed discussion of this process may be found in Tomasello (2003).
English. In her opinion, as the child acquires more and more formulas, the representations of newly acquired formulas overlap with those that have been acquired earlier, which leads to the formation of generalization. For example, let us assume that the child has learnt three multi-word units: Where’s the ball?, Where’s daddy?, and Where’s the milk?, analyzing them in the following way: ?PLACE + THE BALL, ?PLACE + DADDY, and ?PLACE + THE MILK. When these three partially analysed formulas are superimposed one on another, a specific schema emerges. The structure of this schema will be the following: the elements that these three formulas have in common, i.e. ?PLACE will remain unchanged, whereas the elements not shared by the formulas will create their less specific, i.e. more general representation [THING/______ ]. Evidently, it is not meant here that mental representations of the formulas overlap in a literal sense. Nonetheless, there exists tangible evidence, bearing out the fact that in the human brain specific populations of neurons represent similar conceptualizations (Singer (2000)). The cognitive theories presented above have been the most influential recently and have had far-reaching effects for the field of first language acquisition. Nevertheless, at this point I would like to present two other cognitive approaches, perhaps less popular, but still interesting with respect to their main assumptions. 1.5.3.3. Other cognitive theories of first language acquisition
In this section I would like to present two other cognitive theories of first language acquisition, i.e. the connectionist model and the information processing model. The connectionist, once referred to as the associationist model, was widely adopted in the 1980s and 1990s to account for language learning in general as well as L1 acquisition. In this approach, the brain is compared to a computer consisting of neural networks, which would be complex clusters of links between information nodes. These connections get stronger or weaker depending on whether they are activated frequently enough or not. According to this theory, learning is possible thanks to associative processes, not the construction of abstract rules. In other words, the human mind is endowed with the ability to look for associations between particular elements and create links between them. As the associations keep reappearing, these links are strengthened, until finally they become part of larger networks as the number of connections between particular elements gradually increases. In terms of language acquisition, this means that both learners of a foreign language and children acquiring their native language are sensitive to regularities in the language input and are predisposed to pick out the most
43
probabilistic patterns on the basis of their regular recurrence in a particular linguistic environment. The process of learning or acquisition takes place as these patterns become enhanced by repeated activation. As Ellis and Schmidt (1997:153) suggest: Connectionism attempts to develop computationally explicit parallel distributed processing (PDP) models of implicit learning in well-understood, constrained, and controllable experimental learning environments. The models allow the assessment of just how much of language acquisition can be done by extraction of probabilistic patterns of grammatical and morphological regularities. Because the only relation in connectionist models is strength of associations between nodes, they are excellent modelling media in which to investigate the formation of associations as a result of exposure to language. Evidently, this approach differs considerably form all the generative theories because it suggests that language acquisition takes place due to the associative processes that have just been described. In this model, then, language is not rulegoverned and is not regarded as a set of syntactic, morphological and phonological rules, accompanied by the lexicon. According to this theory, learning is not based on extracting any rules by children from the language around them so as to establish their own set of rules but it involves the construction of associative patterns. Connectionism in that view is seen as an alternative to symbolic accounts of language acquisition: rule-like behaviour does not imply rule-governed behaviour (Ellis (1996b:364)). When it comes to the application of this model to the studies of L1 acquisition, researchers have devised computer programmes designed to build the kind of neural networks which are set up in human mind during the process of language acquisition. These models create a certain network from the linguistic or other input that they receive and the output of the model is compared to the natural – that is human – output. Let us now provide an example illustrating this procedure, taken from the pioneering work of Rumelhart and MacClelland (1986) who designed a computer model simulating the learning of the regular versus irregular past tense in English, based on associative patterns. It is commonly known that children when acquiring the irregular past tense in English, go through three phases (Mitchell (1998)). First of all, they produce irregular past tense forms correctly (e.g. ate, came), then they make an overgeneralization of the regular past tense ending to irregular forms (e.g. eated, falled) and in the third and final phase they produce irregular forms correctly again. This pattern is accounted for by claiming that children begin with rotelearning a few common irregular past tense forms since many of the common verbs in the early child language are irregular (e.g. go, eat, give, come, throw, 44
fall, take, drink, etc.). Only later on do they extract from the linguistic input the rule stating that most commonly past tense is produced by means of -ed ending. Being unable to allow exceptions at that stage, they overgeneralise this general rule to all verbs indiscriminately. The last stage in this process involves disposing of the overgeneralization of marking, which leads to the emergence of the system acceptable in the adult language. This model is referred to as the Ushaped curve of learning. Rumelhart and McClelland’s learning model turned out to bear great resemblance to the way in which children acquire past tense in English because the computer made similar generalizations and overgeneralizations on the basis of the provided input. This model has given rise to many studies in recent years (e.g. MacWhinney and Leinbach (1991) Plunkett and Marchand (1991)). Another cognitive model of first language acquisition is the so called information processing model. Information processing models were developed by cognitive psychologists and only later were they adapted to the treatment of language processing, both L1 and L2. I am going to concentrate on McLaughlin’s model as it is fundamental to this theory and, furthermore, it inspired a great deal of other studies. According to McLaughlin, [...] the fundamental notion of the information-processing approach to psychological inquiry is that complex behaviour builds on simple processes (McLaughlin and Heredia (1996: 213)). Moreover, these processes are modular, which means that they can be studied independently of one another. Here is the summary of the main characteristics of this approach (Mitchell (1998)): 1. Humans are viewed as autonomous and active. 2. The mind is a general-purpose, symbol-processing system. 3. Complex behaviour is composed of simpler processes. These processes are modular. 4. Component processes can be isolated and studied independently of other processes. 5. Processes take time; therefore, predictions about reaction time can be made. 6. The mind is a limited capacity processor. Thus, within this framework to acquire a language is to acquire a complex cognitive skill which consists in the automatization of component sub-skills needed to perform a given linguistic task. To sum up, cognitive theories of language acquisition emerged as a result of the growing disappointment with the theory of generative grammar which proved incapable of accounting for several phenomena of the first language acquisition process. The origin of cognitive theories can be traced back to the semantic approaches that inspired a great deal of research in the field. In the 45
light of cognitive linguistics, in the initial stages of linguistic development the process of language acquisition consists in mastering specific constructions built around lexical elements, whereas the role of linguistic creativity is considered to be minor. Reaching more advanced stages of cognitive and linguistic development, children become more creative, as they abstract symbolic units of language, i.e. schemas from fixed phrases, i.e. formulas that they have acquired.
Conclusion On the whole, the approaches to the first language acquisition described in this chapter apparently represent the everlasting dichotomy of views concerning the question of the specificity and innateness of the language faculty, which is far from being resolved. As Butterworth and Harris (1994:124) have suggested: In some respects both the claims of Chomsky’s and Piaget are correct. There is evidence that acquisition of some aspects of language, notably syntax, is independent of other aspects of cognitive development.... At the same time, however, there is no doubt that full understanding of a great deal of language requires other, more general cognitive abilities. Nonetheless, Chomsky’s theory of UG is nowadays in retreat, as the majority of the first language acquisition researchers have adopted the cognitive framework for the child language analysis and description. As recently more and more research has been done in the field of first language acquisition, it has turned out that many of Chomsky’s assumptions cannot be verified empirically. Moreover, the strict formality and high degree of abstractness of UG theory made linguists turn to semantic approaches to first language acquisition which, with the passage of time, evolved into cognitive approaches. In subsequent parts of this book I will rely mainly on semantic and cognitive approaches to first language acquisition because, as my research has shown, the early child speech is evidently constructed around lexical elements. What is more, it is inherently linked with the child’s cognitive development and cognitive abilities responsible for its organization. In the next chapter I will prove that the earliest stages of language acquisition consist in mastering specific semantic notions and semantic relations. Morphological and syntactic rules are virtually non-existent at this stage. Therefore, the child acquires entire constructions from which only with the passage of time does he abstract formulas, and more abstract schemas.
46
CHAPTER II
Omission of content and function words
It is good to express a matter in two ways simultaneously so as to give it both a right foot and a left. Truth can stand on one leg, to be sure; but with two it can walk and get about. (Friedrich Nietzsche, quoted in Hollingsdale, J.R. (1999:54))
Introduction The aim of the present chapter is to present a wide array of formulas, i.e. fixed phrases, acquired by children, from which, according to cognitive linguistics (Dąbrowska (2003)), symbolic units, i.e. abstract schemas, are derived. In the light of traditional theories of first language acquisition the initial stages of linguistic development consist in mastering ‘key’ lexical elements, disregarding ‘less important’ lexemes and grammatical morphemes. In what follows I will try to combine these two approaches. Therefore, the core of the present chapter is an attempt to discuss the omission of various sentence elements that can be observed in the child language in the early stages of its development. I will provide a detailed analysis of this phenomenon, examining the whole spectrum of such incomplete sentences from the language of children in the light of the phenomenon of language redundancy. For that purpose I will rely both on traditional and cognitive studies of first language acquisition. To begin with, I will provide a definition of language redundancy, illustrating it with relevant examples drawn both from child and adult language. Next, I will have a look at the explanations that have been proposed for the omission of both function and content words in child language and I will verify these theoretical considerations on the basis of relevant data drawn from Polish and English. I will dwell on the question of the ambiguity of telegraphic utterances, trying to provide their grammatical and semantic interpretation. My considerations concerning the language of children will be based almost entirely on my own language corpus, gathered in the course of longitudinal studies of 47
three Polish and three English-speaking children. The data was collected mainly in the form of audio recordings for three years, from the year 2000 to 2003. All the children were recorded regularly for periods of time ranging from a few months to two years.14 When giving the examples of their speech samples from my corpus, I have also decided to include the age of each child in brackets next to the letter standing for its name. Thus, E (1;8) I’m sleep means that the utterance comes from Ella, who was 1;8 at that time. As regards language redundancy, first I will discuss this phenomenon taking a broader perspective, i.e. considering its various types and its consequences for the process of communication. Then I will concentrate on language redundancy in child language, introducing, for that purpose the term telegraphic speech.
2.1. Language redundancy Each language has built-in redundancy; therefore, it can be understood at least to some degree even if, for some reason, its utterances are deprived of either entire sentence constituents or only the grammatical morphemes required to generate grammatically well-formed sentences and, at the same time, encode semantic modulations. Thus, linguistic redundancy functions as a back-up system in order to sustain communication in the case of imperfections and failures occurring during the communication process. In other words, […] redundancy is a means of keeping the system running in the presence of malfunction. (Campbell (1982: 73)) The issue of redundancy has been dealt with over the past several decades and, as a result, several definitions of this phenomenon have emerged. For example, Hunnicut (1985) has stated that […] redundancy is a feature of an information source which insures that the communication receiver is able to reconstruct a message that has somehow suffered from transmission interference or deletion, and so interpret it satisfactorily. Another definition has been formulated by Campbell (1982: 68), who claims that […] extra ration of predictability is called redundancy. It means that language users are able to predict and reconstruct the missing bits of information. Making linguistic predictions is possible thanks to the fact that language is systematic and rulegoverned. As Hunnicut (1985: 53) has put it: […] redundancy is the systematicity in one’s language and speech. For the purpose of my further considerations I am going to discuss the model of Wit and Gillette (1998) who distinguish between grammatical and 14
Detailed information concerning the procedure of compiling the corpus has been included in the introduction.
48
contextual redundancy. In my work I am primarily concerned with grammatical redundancy, by which they mean […] the internal systematicity and rulegoverned behaviour of a language in which two or more of its features serve the same function. These features are obligatory and may not be omitted. Grammatical redundancy [...] is internal to the language in the sense that it is generated from grammatical rules and independent of situational, contextual and non-linguistic considerations (Wit and Gillette (1998: 3)). An example of this is the English third person singular morpheme -s, which, together with the subject of the sentence, provides information about the doer of the action. Since English is not a pro-drop language, the presence of the subject is obligatory, thus making the -s morpheme perfectly redundant and, therefore, often omitted both in the course of L1 and L2 acquisition. Unlike grammatical redundancy, contextual redundancy is voluntary because its use depends on the judgement of the speaker concerning the receptor’s background (e.g. whether he is familiar with the topic) and it may be resorted to so as to achieve a certain rhetorical effect, e.g. to enhance comprehensibility, resolve ambiguity, contrast some elements or emphasise the point. This is achieved by […] the repetition of information that is, in a grammatical sense, non-obligatory. This repetition consists of the reproduction of identical elements of information or of elements that are only apparently identical (Wit and Gillette (1998: 6)). For instance, the sentence: She has not seen ’Beautiful Mind’, the Oscar winner contains two pieces of information, the latter of which is redundant for any film fan since it is well-known that the film won an Oscar. Thus, in this context, the repetition of information serves the purpose of intensifying the meaning of the expression: She has not seen ’Beautiful Mind’ although it is a good film because it has been awarded with an Oscar. However, if I assume or know that my interlocutor is not interested in the cinema and not aware of the first-rate quality of the film, the tag the Oscar winner is not redundant because it provides the listener with the new information, which is not, by any means, repetitive. Thus, contextual redundancy is context-dependent because the same information in a particular sentence may be perceived either as redundant or basic. The same kind of redundancy may be observed in L1 acquisition: K (2;4) Babcia mi dała dziecko nowe, lalę, dzidziuś15 ‘Granma has given me a new child, a doll a baby’, where it is used for clarification,16 i.e. to specify precisely what the child got from her granma. 15
The nominative form dzidziuś ‘baby’ has been used here instead of the genitive case dzidziusia because of the child’s insufficient command of the Polish morphological system. 16 This phenomenon is an example of how redundancy may be used purposefully as a rhetorical device. Nevertheless, this issue is beyond the scope of my considerations since the focus of my
49
However, for the purpose of my research these two definitions of redundancy are not sufficient because they do not account for a wide variety of phenomena in children’s speech. Although grammatical redundancy provides an explanation for the ability to understand sentences deprived of grammatical morphemes, there is nothing to rely on when trying to analyse the omission of entire sentence elements, such as subject, object, verb, etc. For this reason, I am going to introduce the notion of pragmatic redundancy, by which I will understand providing information, necessary from a grammatical and logical point of view that, however, could be inferred on the basis of the overall, extra-linguistic context of the conversation, i.e., people and things in the child’s vicinity that he may refer to, or his previous linguistic behaviour in a similar situation. Moreover, a lot can be inferred from the gestures and body language of the child. Thanks to such situational clues, the caretakers can infer their children’s intentions, wishes, desires, fears, etc. in particular circumstances, even if the child’s utterances are deprived of many elements crucial for their understanding, such as subject, object, verb, attribute, etc. In other words, these situational clues function as a back-up system […] in the presence of malfunction. Another very important factor facilitating comprehension of incomplete child language is of a psycholinguistic nature, i.e. the fact that parents know the psychological traits of their children very well and also their way of expressing themselves and formulating their thoughts as specific language users. Thanks to this knowledge, they deduce the meanings intended by children by expecting, i.e. predicting what they want to say, because, as I have already remarked, […] an extra ration of predictability is crucial for the existence of redundancy. These considerations can be exemplified by one of the sentences from the corpus, where E.’s (1;4) utterance Waltz was meant to express the following meaning: ‘I want to dance a waltz’. The reconstruction of this message was possible thanks to situational clues, such as pointing to the relevant record, looking in the direction of a CD player, etc. It was also the result of the previous parental experience because, in the past, whenever the girl uttered this word, it was always equivalent to her willingness to dance. However, let us imagine the situation in which the word waltz is used by the same child to express two different intentions: to ask to dance a waltz and to listen to it. Then, whenever this word is uttered again, we face the task of working out which proposition it stands for, i.e. its meaning is ambiguous. Thus, the existence of linguistic redundancy, of whatever kind, entails the ambiguity of the language because all its incomplete sentences give rise to their numerous interpretations. interest will be a mere existence of redundancy in the language and its immensely important role for interpreting children’s utterances from which some sentence elements have been deleted.
50
Because of this, the existence of redundancy has serious implications for the communication process, as it may cause some misunderstandings or confusion if a particular message is interpreted against the speaker’s intentions. On the other hand, owing to this phenomenon, speakers using the same language are capable of communicating with each other even though their command of this language is far from being perfect. Although such a process is possible, it is full of imperfections and flaws, as a multitude of grammatical and, at the same time, semantic relations escape being articulated. This in turn results in great ambiguity because utterances deprived of entire constituents or consisting of words lacking grammatical morphemes are susceptible to a variety of interpretations and can always be built up into a number of different sentences. The phenomenon of language redundancy is particularly conspicuous in the speech of children acquiring their mother tongue. The developmental restrictions and processing limitations of the children’s cognitive system enable them to generate nothing more than simplified utterances lacking lots of important elements. Studies of early child speech have shown that young children tend to omit words of low information value (having little meaning on their own), such as function words: copulas, articles, prepositions, auxiliaries and also inflectional morphemes, i.e. the same sort of words which adults leave out in their telegrams. For this reason, the language of children at an early stage is referred to as telegraphic speech (Brown 1973:74). A typical example of the telegraphic speech would be the following sentence: J (1;8) Train gone for ‘The train has gone’ or K (2;4) Stałam drzwiami ‘I was standing door’ instead of Stałam pod drzwiami ‘I was standing at the door’.17 However, when looking at the early stages of children’s speech more closely, it turns out that it is not only function words that children omit but also a great many content words (which have lexical meanings). In this case, the criterion of what can be omitted is much more subjective because it is applied in accordance with the child’s inner feeling of what is important enough and therefore worth talking about and, on the other hand, what is less important, or obvious and consequently can be left out. To illustrate this phenomenon I would like to give an example from my longitudinal studies of Polish children. When Z (1;11) was drawing with a crayon, she said Zuzia złamała ‘Zuzia broke’, which was supposed to mean ‘Zuzia złamała kredkę’ ‘Zuzia broke the crayon’. In this way, the girl wanted to point to the action performed on the object, not the object itself, which was obvious from the situational context and therefore not mentioned. 17
The examples presented in this chapter are mine, unless another source has been given.
51
I have decided to include the utterances of this type in the present discussion although they are not examples of telegraphic speech. However, they are also shortened, i.e. incomplete versions of messages, i.e. the telegrams that the child is trying to communicate to the outside world, containing only the key words, from the child’s point of view, and lacking these vocabulary items that can be inferred from the context.
2.2. Definition of telegraphic speech In this section I will focus on the criteria that an utterance must fulfil in order to be regarded as an example of telegraphic speech. Then I will go on to discuss the characteristic features of telegraphese from a cross-linguistic perspective. The period of telegraphic speech begins at the two-word stage which, as language acquisition experts postulate, extends approximately from the eighteenth up to the twenty fifth or even to the thirtieth month of the child’s life. It is then that many remarkable linguistic changes take place. Especially marked are the changes in the child’s vocabulary which increases to hundreds of items. This rapid growth in the size of active vocabulary entails another important development, namely the child’s attempts to put together newly acquired words. The result of these attempts are simplified sentences characterised by the lack of function words, grammatical morphemes and contentives. However, unlike telegrams, which also omit certain words, the children’s telegraphese is not purposeful but results from developmental restrictions. The beginning of the two-word stage marks the onset of telegraphic speech, which continues well until children master basic syntactic rules. This usually takes place at around the age of three, when they have already learnt to produce all the necessary sentence constituents. In my discussion I will devote some attention to this phenomenon not just during the two-word stage but throughout the whole period of acquiring the basics of syntax. We have stated that the period of telegraphic speech commences at the onset of the two-word stage. However, during the two-word stage the language of children contains many one-word utterances. The question then arises: is it justifiable to classify single words as examples of telegraphic speech, expressing entire propositions? Brown (1973) believes that single words during Stage I18 are not used at random, but are parts of highly organised units since the child 18 Single words before Stage I, i.e. before any word combinations enter into the child’s speech, are used without any grammatical knowledge because this knowledge has not yet become available to the child.
52
already possesses some rudimentary knowledge of how the sentences are structured and uses words as parts of sentences, assigning to them functions of agents, objects, etc. (at the same time sometimes failing to produce other necessary constituents). As regards the general characteristics of telegraphic speech formulated on the basis of data drawn from Polish and English, they are as follows: 1. Children’s telegraphic utterances contain mainly nouns and verbs, as these are the parts of speech that the child acquires first. This cognitive and linguistic set-up is deeply rooted in the sensori-motor stage during which the child learns to understand his environment mainly by acting on it. By using his senses of touch and sight he begins to perceive the world both in terms of what surrounds him, i.e. things, and also relationships affecting objects as well as actions performed on them, i.e. states and actions. 2. In English, word order is preserved in the vast majority of cases. In Polish, word order varies because it is used to point to new information.19 However, 19
To realise the importance of word order in the two languages, it is essential to know that English is referred to as a non-inflectional, grammatical word order language, while Polish is classified as an inflectional and pragmatic word order language (Fisiak, 1978). Grammatical word order languages encode grammatical relations on word order patterns because their systems of inflectional morphology are impoverished. With regard to English, the normal, i.e., unmarked, structure of the sentence is SVO, influenced little by pragmatic factors. Any changes in the word order would result in either ungrammaticality (see 1b, c, d) or a change in meaning (e.g. 2b) of the sentence: 1a. George bought a ring for his wife. 1b. *For his wife George bought a ring. 1c. *A ring George bought for his wife. 1d. *George bought for his wife a ring. 2a. The ball hit the wall. 2b. The wall hit the ball. Therefore, the elements to be emphasised in the sentences are highlighted by means of stress or structural devices, such as the passive voice or cleft sentences. In contrast, in Polish syntactic functions are marked by case endings and, basically, they are not affected by changes in the word order. There exists only one case where position functions as a marker of a syntactic function. This is when subject and object have identical case endings, which is possible because in some declensions nouns have the same form in the Nominative and Accusative case, e.g. Szczenię pogryzło cielę. (The puppy bit the calf.) In this case a rearrangement of subject and object results in a substantial change in meaning: Cielę pogryzło szczenię. (The calf bit the puppy.) Therefore, if I examine the Polish counterpart of sentence 1a to which the same changes as in 1b, 1c and 1d have been made 3a. Jerzy kupił pierścionek swojej żonie. 3b. Swojej żonie Jerzy kupił pierścionek. 3c. Pierścionek Jerzy kupił swojej żonie. 3d. Jerzy kupił swojej żonie pierścionek.
53
especially in the later stages of the children’s linguistic development, appropriate inflections are used, which provides evidence for the fact that in both languages the internal structure of the sentences is preserved. 3. Telegraphic speech is observable both in imitations and spontaneous sentences (Brown and Frazer (1964), Brown and Bellugi (1964)). Thus, in the light of traditional linguistics, telegraphic speech is the first stage in the overall process of speech development in which children try to combine words with a view to forming sentences. Although these sentences are in many ways deficient, they preserve the deep structure of the language being acquired and ensure gradual transition from the one-word stage to the stage of multi-word utterances. In contrast, according to cognitive linguistics (Dąbrowska and Kubiński (2003)), during the telegraphic speech stage children acquire nothing more than incomplete phrases constructed around lexical not grammatical elements.20
2.3. Explanation of omissions When talking about possible explanations for the omissions, I must differentiate between two different kinds of omissions: deleting grammatical morphemes and leaving out entire sentence constituents, such as object, agent, action, etc. In the study of Brown and Frazer (1963) which consisted in eliciting imitations of 13 simple English sentences, function words were more often omitted than content words. From my data it is clearly visible that in spontaneous productions, both in English and Polish, in the initial stages, the omission of content words is more prevalent due to the primitive syntax, i.e. inability to form sentences or even longer phrases. In the light of cognitive studies of first language acquisition young children do not possess any abstract syntactic knowledge (Tomasello (1999)). According to Tomasello, children until we can see that all the sentences are not only grammatical but also have the same meaning and the only function of word order is to express special emphasis on certain elements. Thus, sentence 3a is an example of the standard, unmarked SVO order, whereas sentences 3b, c, and d are contextually marked, i.e. the words expressing information considered to be the most important or new in a given context have been highlighted by changing their position in the sentence. Basic structural differences between English and Polish can be summarised in the following way: Those languages which encode more pragmatic distinctions explicitly will exhibit correspondingly less pragmatic ambiguity [...] I would expect the more extensive pragmatic use of word order to correlate quite generally across languages with a richer morphology, with greater surface structure disambiguation, less semantically diverse grammatical relations, less raising, fewer distractions and fewer deletions, while the more grammatical use of word order will correlate with just the reverse (Hawkins 1983:125). 20 This particular issue will be elaborated on throughout the following subchapters.
54
the age of between 2;0 and 2;6, apart from the categorisation of nominal elements, are incapable of creating and using any abstract linguistic units, or forms of syntactic organisation. Only as late as at the age of three do children begin to form abstract categories. Consequently, till this age their speech is not based on syntactic knowledge but is organised around specific lexical items. Therefore, the early utterances of young children are fragmentary and incomplete with many missing elements. However, as the children’s speech develops, i.e. as children start to acquire abstract syntactic knowledge, their language begins to contain more and more sentence constituents, and consequently it is then that functors are more often omitted than contentives. 2.3.1. Omission of functors The function words that are left out have little phonetic substance, i.e. are not perceptually salient (Blaisdell and Jensen (1970)) and express minor semantic roles (Scholes (1969, 1970)). Blaisdell and Jensen (1970) found that both primary stress, final position and a syllabic nature favours reproduction of the functor by the children. Another variable affecting the ease of acquisition is whether or not the morpheme is phonologically conditioned by the verbal context, i.e. the phonological properties of the verb stem. If it is, as is the case with the majority of English morphemes, it is much harder to acquire because then the child faces the task of mastering a few allomorphs. Scholes (1969) has suggested that if a functor or a contentive expresses an important semantic role, it will be retained. Brown (1973) has proposed a more detailed processing of variables accounting for the acquisition of functors in Stage I speech on the basis of Slobin’s (1971) operating principles. They are as follows: 1. If functor x is characterised by low frequency and high perceptual salience and expresses a basic semantic role not a modulation, it will be used freely and correctly in Stage I. 2. If functor x has high frequency and high perceptual salience, whatever its semantic role, it will occur in Stage I but only in prefabricated routines.21 21
Prefabricated routines are created when functors are incorporated into words preceding them due to the inappropriate segmentation of the input, i.e. because of the assumption that the functors are word final sounds of the antecedent words, especially if t hey follow these words without any pause or other juncture sign. These – from a grammatical point of view – overextensions may pass unnoticed in contexts which require the two morphemes, for instance: This a book or I go get it. Only ill-formed sentences are revealing in this respect, e.g., Have a-pants, Mommy get-it ladder, It’s went (R. Clark; 1974). A detailed discussion of rote-learned routines may be found in Mac Whiney (1982, 1985).
55
3. If functor x has low frequency and low perceptual salience and expresses merely a semantic modulation, it will be completely absent from Stage I. In general, cross-linguistic studies of language acquisition confirm a palpable connection between the frequency, perceptual salience and semantic role of a given functor and its acquisition. For instance, in Polish asyllabic prepositions (w ‘in’, z ‘with’) tend to be omitted more often than syllabic ones (do ‘to’, na ‘on’).22 In English, the uncontractible copula is acquired much earlier than the contractible, which is always missing until 2;0 (Brown (1973: 278)). Somewhat contradictory evidence has been provided by Brown and Fraser (1963), who have shown that pronouns, despite their high frequency, and perceptual salience (they are syllabic and can be stressed) do not occur only in prefabricated routines. On the contrary, they were correctly imitated 72 per cent of the time, whereas for inflections the figure was 44 percent and for articles 39 per cent. The question then arises: what is the reason for this phenomenon? The answer may be that the correct usage of pronouns is closely related to whether or not they are often used in fixed phrases in the carer speech. This means that if the child often hears sentences such as Have it or Get it, he will treat them as unanalysed wholes and produce the pronoun only in prefabricated routines. In the light of cognitive theories of first language acquisition, children acquire mainly fixed phrases at the beginning of the multi-word stage. For example, Lieven et al. (1997) has concluded that as many as 92 per cent of all the children’s earliest multiword utterances are made of fixed phrases. That is one of the reasons for the high percentage of pronouns which occur in many stereotypical constructions. Moreover, if pronouns occur often in novel utterances, never heard before by the child and if mentioning them is important for the meaning of the whole sentence, they will not be deleted. For example, in the sentence K (2;4) Daj mi ‘Give me’, the object pronoun mi ‘me’ is used here in the major semantic role, i.e. that of a patient, which means that it is crucial for the understanding of the sentence and therefore it has been retained. Apart from this, pronouns function in other major semantic roles, such as agent, benefactory, etc. Another thing is that pronouns tend to be omitted more often if they function as agents since it is assumed that the doer of the action is always known because it is either the child itself or it is obvious from the context who performed the action. For instance, J.(3;0), while looking at the pictures in his fairy tale book said: Bumped his head ‘The cat bumped his head’, assuming that everybody knew that he was talking about the cat depicted in the picture. 22
56
See Smoczyńska (1985).
Interesting evidence has been provided by Park (1970a) in his study of the acquisition of German. He found that Ulrike and his other children at Stage I produced or imitated certain functors as often as they did contentives. Park’s utterances from his children include numerous occurrences of the following function words: das ‘the’, ein ‘a’, mehr ‘more’, andre ‘another’, hier ‘here’ as well as separable verb prefixes: ab ‘away’ or ‘from’, an ‘on’, ‘onto’ or ‘at’, mit ‘with’ and some others. As far as using indefinite and definite articles, such as ein and das is concerned, the sample was too small to ascertain that it can be representative for German children at Stage I. However, in the case of other function words there is a lot of parallel evidence from Polish and English that they do occur in children’s speech. Their presence can be justified mainly on semantic grounds. First of all, here and there may operate as demonstrative pronouns both in English and Polish, e.g. Here book or Tam misiu ‘There teddy-bear’ when the child refers to a given object in his environment in terms of whether it is near to or distant from him. Secondly, here and there can function as pro-locatives (Brown (1973: 81)), i.e. forms replacing full locative phrases, e.g. K (2;3) Kubek tutaj ‘Mug here’, which was supposed to mean: ‘Postaw kubek na stole’, i.e. ‘Put the mug on the table’, where tutaj ‘here’ stands for a full locative phrase na stole ‘on the table’. In my data the occurrence of these two functors as prolocatives is extremely frequent because they are one-word phrases, much simpler to produce than complex locative phrases. In the light of cognitive linguistics, the ability to point not only to places, but also to people, things or the time is possible thanks the rule of indexicality in language (Tabakowska (2001)).23 The words that are used for pointing are called deictic expressions: here, there, now, then, this, that, I, you, he, we, etc. The use of deictic expressions, which is very common in children’s speech, makes the adult listener perceive the world from the child’s point of view. Thus, here and there indirectly mark the child’s position in the world around him. This phenomenon, referred to as deictic orientation, is responsible for establishing the speaker’s position in space and time. Since one of the characteristic features of children’s speech is to perceive the world from their own, egocentric point of view (Bowerman (2003)), prolocatives functioning as deictic expressions are common in their speech. Secondly, more and another, quite prevalent in English, and their Polish counterparts, więcej and drugi, inny express a basic semantic relation for Stage I, namely recurrence. To be more precise, these words comment on or request for the recurrence of a referent e.g. Z (1;9) Drugie serduszko ‘Another heart’, which 23 The rule of indexicality is based on the existence of an index, i.e. a sign whose function is to point to things, objects, people, phenomena, etc. A detailed discussion of indices and the rule of indexicality may be found in chapter III.
57
is a telegraphic version of the imperative: ‘Narysuj mi drugie (jeszcze jedno) serduszko’ ‘Draw another heart for me’. Another example would be K (2;0) Drugie ‘Another’, this time being a comment on the recurrence, supposed to mean ‘Drugie światło też świeci’ ‘Another light is also on’. Finally, studies of the acquisition of American English have confirmed Park’s (Braine (1963)) observation that verb particles sometimes function as the names of actions, replacing the full verbs to which they belong and thus expressing basic semantic relations, not just semantic modulations. The examples recorded by Braine are Boat off and Water off. The example from my data would be Out for ‘I will get it out’. Summing up, the semantics of the functors seems to be the key factor in their acquisition, far more important than their perceptual salience, whereas the significance of frequency needs yet to be verified in the face of the lack of uniform evidence either in favour of or against its impact on the whole process. Apart from the facilitating determinants of the acquisition of functors, there are also those that have an inhibiting role. According to Brown (1973: 83), [...] redundancy militates against acquisition of a functor. Consequently, those functors, either bound or free, which are perfectly predictable from the context and therefore redundant, will be acquired later than ones which are not predictable and have an informative value. For instance, the third person present indicative inflection in English can be easily inferred if the subject is expressed, therefore it is considered to be redundant and acquired by English-speaking children relatively late (in Brown’s order of the acquisition of the fourteen basic grammatical morphemes in English it occupies the tenth place). On the other hand, the prepositions in and on are neither predictable from the context nor expressive of semantic modulations but they do have basic semantic roles and for that reason they are acquired relatively early (according to Brown, they follow directly after the present progressive inflection -ing, which is acquired first). Having briefly discussed the conditions conducive to retaining or deleting functors, let us now focus on another aspect of omissions in children’s speech, namely that of leaving out contentives. 2.3.2. Omission of contentives There exists a multitude of theories accounting for the delayed emergence of major constituents of a sentence, such as agent, action or object. The least interesting from the linguistic point of view is the case when the child does not know the relevant word. However, it is often the case that children have already acquired lexical items needed to express semantic roles, but still delete them. The question, then, arises: what makes children omit certain parts of the sentences, even though they are familiar with the words needed to denote them? 58
One of the most important reasons for this is of an extra-linguistic nature, mentioned already in the introduction. In the world of the Stage I child, limited, in large part, to his home (i.e. the familiar setting) and his mother (i.e. his main interlocutor, who knows her child very well, and who is capable of guessing much of the information absent from her child’s speech), the needs of communication rarely require all sentence constituents obligatory in the adult language: The familiar setting, the shared memories, the tendency to be concerned only with the present go far toward making the speech redundant (Brown (1973: 205)). Sheer linguistic prerequisites are discussed in sections: 2.3.2.1 and 2.3.2.2. 2.3.2.1. Overgeneralisations
Roughly speaking, the omission of entire sentence constituents, i.e. content words, can be a result of overgeneralisations that children make about the way in which language is structured (de Villers and de Villers (1985)). This phenomenon is also responsible for the omission of some functors, especially auxiliaries and copulas. When models for children’s utterances, i.e. parental performance are considered, it can be easily observed that apart from complete sentences they also include numerous non-sentence fragments that are usually well-formed constituents of sentences (Brown (1973)). For example, in colloquial English and Polish it is possible to provide answers to specific questions, consisting of just those constituents that are required to be specified. Thus, the examples of this phenomenon derived from English would be the following: Who dressed the doll? What did she do? What did she dress? How did she dress it? and from Polish, respectively:
Julie. Dressed the doll. The doll. Nicely.
Kto ubrał lalkę? Co ona zrobiła?
Julia. Ubrała lalkę.24
24
In Polish ellipsis is more easily allowed than in English in both affirmative sentences and questions (Fisiak (1978)). As regards affirmative sentences, in the majority of cases the subject may be deleted because it is encoded in the inflectional morpheme of the verb stem. For example, in the verb poszli ‘they went’ the ending -li denotes the masculine and plural form of the noun or pronoun representing the doer of the action but absent form the sentence. Therefore, it is acceptable to say: Oni poszli na zakupy ‘They went shopping’
59
Co ona ubrała? Jak ją ubrała?
Lalkę. Ładnie.
Furthermore, in spoken language there are some other rules permitting nonsentence fragments. 1. Adults often produce complete declarative sentences and then after a while, as a result of an afterthought, they add a new constituent, which is not a part of this sentence but a separate, independent utterance. For instance: That’s Daddy coming. From work.
To tata wraca. Z pracy.
2. Sometimes unexpressed constituents are obvious both to the speaker and listener from the context and in spoken discourse they can be omitted. Thus, when a child is drawing, its mother can say one of the following: Nice. Another pencil? Finished?
‘That’s nice.’ ‘Do you want another pencil?’ ‘Have you finished?’
Ładne. Inną kredkę? Skończyłaś?
‘To jest ładne.’ ‘Czy chcesz inną kredkę?’ ‘Czy już skończyłaś?’
as well as: Poszli na zakupy ‘Went shopping’. For this reason, the Polish sentence: Ubrała lalkę is correct even in the formal variety of Polish, whereas its English counterpart: ‘Dressed the doll’ may be considered acceptable only as an answer to the specific question in colloquial English. On the other hand, when referring to the thing or object for the first time in Polish, the subject must be mentioned, otherwise the meaning of the sentence would be vague. For instance, the utterance of K (2;3): Nie jedzie ‘Is not going’. is completely unclear since the agent had not been referred to before and what is more, could not be inferred from the context at the moment of speaking. In English the subject is an obligatory constituent of every clause as verbs lack the inflectional endings and, consequently, provide no basis for decoding the doer of the action. The only exception is the imperative clause where the subject is left out.
60
3. In English,25 in colloquial speech it is acceptable to form truncated Yes-no questions, in which both the subject and auxiliary can be omitted: Can you see this truck? You see this truck? See this truck? 4. In addition, parental performance in both languages includes many imperatives, e.g. Umyj ręce ‘Wash your hands’, which gives children the impression that subjectless sentences are more common than they really are. All in all, it probably seems to the child that ellipsis is more freely allowed than it really is, and that both a full sentence or just any part of it can be produced in a discourse. In other words, children overgeneralise the rules of ellipsis, repetition and colloquialism and produce a great many sentences deprived of their essential constituents: 1. They form subjectless sentences when the subject is grammatically required: Jk (2;0) Go to work ‘Daddy has gone to work’, J (3;0) Don’t know for ‘I don’t know’, J (3;8) Smell good for ‘The dinner smells good’, etc. 25
In general, the structure of Polish questions is completely different from English ones. First of all, no auxiliary is required; secondly word order is relatively free and thirdly it is not necessary to mention the subject if it is clear from the context, or if it is the second person singular: Gdzie pracujesz? ‘Where do you work?’ [Where work?] Yes-no questions are formed by means of the question word czy, and the subject-verb inversion is optional: Czy Piotruś zjadł obiad? ‘Has Peter eaten dinner?’ Czy zjadł Piotruś obiad? However, czy can also be omitted and then the inversion is necessary: Zjadł Piotruś obiad ? ‘Has Peter eaten dinner?’ Otherwise, the question conveys an idea of surprise or astonishment: Piotruś zjadł obiad? Naprawdę? ‘Peter has eaten dinner? Really?’ Nevertheless, the above two questions, irrespective of whether inversion is used or not, can be regarded as complete interrogative sentences since none of the obligatory constituents is missing. Thus, only English truncated questions which lack the auxiliary and the subject can be termed non-sentence fragments and considered to be responsible for the deletions of content words in children’s speech.
61
2. They omit obligatory verbs: Z (1;8) Na koniku ‘On the horsie’ for ‘Jadę na koniku’, i.e. ‘I am riding my horsie’, K (2;2) Żaba piasku ‘Frog sand’ for ‘Żaba wpadła do piasku’, i.e. ‘The frog has fallen into the sand’, etc. 3. They delete objects: Z (1;8) Włącz ‘Turn on’ for ‘Włącz dzidziusia’, i.e. ‘Turn on the baby-doll’, E (1;8) He broke for ‘He broke his leg’, E (1;4) Show me for ‘Show me the shoes’, etc. Cognitive theories of language acquisition account for the acquisition of incomplete language chunks not on the grounds of overgeneralizations made by children, but on the grounds of very limited linguistic creativity in children below the age of three (Lieven (1997), Tomasello (2000), Dąbrowska (2000b)). Lieven et al. have stated that a sweeping majority of the earliest multi-word utterances are made of fixed phrases or stereotypical formulas. In the light of these findings it can be claimed that at the onset of their linguistic development children acquire not single words but whole phrases, and use them in accordance with their communication needs. This leads to situations in which children use just phrases instead of whole sentences; this is why some elements are ‘missing’ from them. However, cognitivists would say that these elements are missing not as a result of overgeneralisations made about the language structure but because of the fact that they have never been there as children learn by rote the whole phrases and use them indiscriminately irrespective of the context. 2.3.2.2. Complexity
Apart from overgeneralisations about the language structure being responsible for deletions, interpreted by the cognitive linguists as the acquisition of entire language chunks, children are also influenced in that respect by the complexity of the sentences. Evidently, the more complex the sentence is, the more difficult it is for the child to produce or reproduce all its constituents. For example, Bloom (1970) has analysed successive two- and three-word utterances in which all elements of an agent-action-object sentence are expressed but never in a single utterance, which suggests that the child finds it difficult to produce all the elements of the more complex structure at one time. She has labelled this phenomenon replacement sequences.26 An example of that would be: 26
According to Braine (1971) utterances of this kind must fulfil one important condition: it is essential that they [...] occur during a fairly short time period during which there is no detectable change in the eliciting situation (i.e. nothing happens in the environment to indicate that the utterances are not equivalent in meaning).
62
Louis read. Read book. Louis book. In fact, in my data there are numerous examples of replacement sequences, which differ slightly from those studied by Bloom because apart from the agent, action and object they also contain a locative phrase, the imperative or the attribute. The reason for this is that they come from children who are linguistically more advanced than those in Bloom’s sample. I have decided to include articulations of this kind in this discussion because they adhere to the same principles, i.e. all the constituents of the utterance are expressed but not in a single sentence, as they should be, but in two successive utterances. Let us now consider examples of replacement sequences from Polish: Z (1;8): Drugi tam. Drugi stoi.
‘Another there’ ‘Another is standing’
Z (1;9): Laluni szynkę Dam laluni.
‘Ham to the doll’. ‘I will give to the doll.’
K (2;2) Kaczka, popatrz Kaczka siedzi
‘The duck, look’ ‘The duck is sitting’
K (2;6) Tam wisieć będzie Karteczka wisieć będzie
‘There will hang’ ‘A slip of paper will hang’
K (2;3) Papierka mam Mały mam
‘I have got a piece of paper’ ‘I have small’
The above replacement sequences constitute yet another piece of evidence in favour of Lieven’s theory (Lieven, et al. (1997)) proposing that in the initial stages of linguistic development children acquire language as a set of fixed phrases and are incapable of combining them with a view to forming one correct and complete sentence. At this point it is interesting to note that the form papierka with the genitive case ending has been used here inappropriately by K., as the correct 63
version of this utterance should be: Papierek mam ‘I have got a piece of paper’, with the noun used in the accusative. This type of mistakes, i.e. using the genitive instead of the accusative, and vice versa can be observed quite often in my data. Let me give a few examples illustrating this phenomenon. First of all, I am going to present sentences in which the accusative has been used instead of the genitive: Z (1;8) Nie jajko (acc.) Jabłko (acc.) nie dał
for
‘Nie ma jajka (gen.)’ ‘There is not an egg’ ‘Jabłka (gen.) nie dał’ ‘He did not give an egg’
Below are examples of sentences in which the genitive has been replaced by the accusative: Z (1;8) Dam27 jogurta (gen.) Daj mi kredek28 (gen.)
for
‘Daj jogurt (acc.)’ ‘Give (me) the yoghurt’ ‘Daj mi kredki (acc.)’ ‘Give me the crayons’
The reason for confusing genitive case endings with accusative ones is that they are used in negative and positive sentences, respectively. Therefore, the child has to realise the existence of two different case endings of the object noun depending on whether the predicate is negated or not. Consequently, it has to be aware of the fact that transforming the sentence from the positive into the negative form or vice versa requires changing the inflectional morpheme attached to the object noun, e.g.
27 It is remarkable that in the first sentence from this set, Dam jogurta ‘Give yoghurt’, apart from using the incorrect case ending, the child also failed to use the imperative form of the verb daj ‘give’, retaining in it the first person singular future form of the verb, i.e. dam ‘I’ll give’, often used by parents when offering something to the child, e.g. Dam Ci jabłko ‘I’ll give you an apple’. This issue will be discussed in detail in the section devoted to morphological development in chapter III. 28 It is disputable whether the form kredek can be at all regarded as ungrammatical because, as Wierzbicka (1988: 448) remarks, the accusative can be used to point to permanence, as in Daj mi nóż (acc.) ‘Give me a knife (for a present)’ while genetive codes temporariness: Daj mi noża (gen) ‘Give mi a knife (for a while)’. Consequently, the imperative sentence Daj mi kredek (gen.) can be interpreted as ‘Give me crayons for a while’. Besides, the genitive can be used in the so called partitive constructions, e.g. Daj mi chleba (gen.) ‘Give me some bread’. When seen in this light, the above utterance Daj mi kredek (gen.) can mean ‘Give me some crayons’. Thus, in Polish the distinction between genitive and accusative is not exclusively of pure grammatical nature as it contains some semantic nuances as well.
64
Mam papierek (acc.) Nie mam papierka (gen.)
‘I have a piece of paper’ ‘I do not have a piece of paper’
Not realising the interrelation between the form of the predicate and the case ending results in using the object nouns with the incorrect inflectional morpheme. An illustration of this phenomenon would be the utterance from this replacement sequence: Mam papierka. Smoczyńska (1985: 640) discusses the issue of confusing the genitive with the accusative case ending as well; however, she claims that it is a unidirectional process. She maintains that children use the accusative instead of the genitive, because when transforming positive sentences into negative ones, they preserve the accusative case ending of the noun object from the positive sentence. Thus, the form of the noun from the positive sentence: Czytam książkę (acc.) ‘I am reading the book’ is retained in the negative sentence: Nie czytam książkę (acc.), whereas the correct version of this sentence would be: Nie czytam książki (gen.) ‘I am not reading the book’. Consequently, she assumes that all children acquiring their native language treat positive sentences as basic, unmarked patterns and they derive other structures from them. Since the children are unable to perform several operations involved in the application of this complex transformational rule, they apply it only partially, i.e. they convert the predicate into the negative form, failing to replace the accusative case ending with the genitive one. Such a line of thought raises two objections. First of all, as has been shown above, it is not only that the accusative is used instead of the genitive in negative sentences: Jabłko (acc.) nie dał ‘He did not give an apple’, but also that the genitive is used instead of the accusative in positive sentences: Daj mi kredek (gen.) ‘Give me the crayons’. Secondly, negation appears in the language of children at the very beginning of their linguistic development, as has been reported by Brown (1973), Bowerman (1973) and Bloom et al. (1975). It has also been observed on the basis of the language corpus that all the six children studied started to incorporate negation into their speech almost as soon as they started to talk, e.g. K (1;8) Kinia nie for ‘Kinia cannot do this’, or E (1;4) No boys for ‘There are no boys here’. Since negation can be considered inherent in the language of even very young children, it is hardly plausible to stick to the idea that utterances containing negation are the result of the transformation of positive sentences. Smoczyńska’s theory that positive sentences constitute the basic pattern is true but only in some communicative contexts, e.g. if the child wants to contradict his parents, he forms the negative sentence by transforming the positive, e.g. 65
Mother: Mamcia ma swoje śniadanko ‘Mummy has got her own breakfast’ Z (2;0): Mamcia nie ma swoje (acc.) for ‘Mummy does not have her own ‘Mamcia nie ma swojego (gen.)’ (breakfast)’ However, this method of forming negations is context-specific and therefore it cannot be said that it is adopted whenever the child wants to make a negative sentence. Moreover, sometimes the opposite is the case, i.e. children convert negative sentences uttered by their parents into positive ones. As an example, I would like to quote the exchange between me and my daughter: I: Nie ma teraz świąt ‘It is not Christmas now’ K (2;8): Jest teraz świąt (gen.) instead of ‘Są teraz święta’ ‘It is Christmas now’ Therefore, in my opinion, one cannot talk about preserving the accusative case in negative sentences, as Smoczyńska has proposed, but rather about mixing up the genitive and accusative case endings in positive and negative sentences, which looks like preserving any case. In the light of cognitive theories of language acquisition the reason for this phenomenon is that complex forms, by which I also understand words with inflectional morphemes attached to them, are first acquired as unanalysable wholes (Langacker (2003)). Consequently, a word that has once been learnt with a particular inflectional ending will be used with this ending for all cases, as children do not realise that in Polish nouns are compositional, and that their inflectional morphemes should be changed depending on the case. Thus, the reason for saying, e.g. Nie czytam książkę for ‘Nie czytam książki’ ‘I am not reading the book’ is that the child acquired the noun książkę as an indivisible and unanalysable unit. Let me now return to the main topic of my considerations and concentrate again on examples of replacement sequences, this time from English: E (1;4) Love you I love E (1;8) Want more tea? Want you have tea?29 29
These two utterances are examples of truncated Yes-no questions, acceptable in spoken English.
66
Bloom (1970) concludes that in replacement sequences all the obligatory constituents are present in the deep structure of the sentence, but they are not expressed in one utterance because of cognitive processing limitations. Apart from that, she has discovered that there are several factors that increase the information processing load and at the same time the likelihood that a given element will be deleted. One such factor is the addition of negation. Bloom (1970) has reported the following sequence: Me like coffee. Daddy like coffee. Louis no coffee. Other factors include the use of two part verbs like turn on or take off, the addition of modifiers to the object noun phrase, e.g. Drive blue car or the use of verbs newly learnt by the child, especially new verbs and prepositions. The use of articles or verb inflections do not contribute to the increase in the number of omissions. Bloom’s conclusions coincide with my findings in this respect. Now I am going to demonstrate that the use of the sentence elements mentioned by Bloom leads to leaving out content words by children both in Polish and in English (for the sake of clarity contentives omitted by children have been written here in bold): a. use of negation: K (2;2) Nie ma tam
for
‘Lali nie ma tam’ ‘There is no doll there’
E (1;4) He not duckie clothes
for
‘He is not wearing the duckie’s clothes’
b. use of a new preposition: K (2;4) Do komatu30 for ‘Idziemy do bankomatu’ ‘We are going to the cash machine’ E (1;4) Round and round
for
‘Wheels go round and round’
30
The deletion of syllables in three-syllable words is a very common phenomenon in the initial stages of the acquisition of both Polish and English. According to Ingram (1992), it is unstressed syllables that are deleted, especially if they precede stressed ones, e.g. E. (1;6) [na:na] for banana, or K.(2.0) [czynka] for dziewczynka ‘girl’.
67
c. use of a new verb: Z (1;8) Gotuje
for
‘Mama gotuje’ ‘Mummy is cooking’
E (1;8) Bouncing
for
‘The monkey is bouncing’
d. use of a modifier: K (2;2) Czyste mam
for
E (2;0) I wanna sharpen the black
‘Mam czyste ręce’ ‘I have got clean hands’
for
‘I wanna sharpen the black pencil’
However, as Maratsos (1984) points out, not all constituents are deleted equally often: for example, deletion of a verb is particularly rare in all English children. As far as my data is concerned, I cannot say that this kind of omission is not that rare but definitely it is quite rare since it accounts for approximately 15% of all the sentence elements omitted. What is more, a specific pattern of these omissions can be detected, i.e. there are two groups of content verbs that are omitted most often: want and can see or can hear. In the utterances in which the child requests to be given a particular object or to be allowed to perform an action, it just mentions the object, the action or the word standing for the action, leaving out the verb want and also the subject I. For example: E (1;2) Up-up More
‘I want to go up’ ‘I want more’
E (1;4) Rock baby Waltz
‘I want to rock my baby’ ‘I want to dance waltz’
E (1;6) Ella fix it
‘Ella wants to fix it’
Secondly, from my data it turns out that children very often tend to draw the listener’s attention to what they can see or hear in their immediate environment. 68
For that purpose in the early period of their linguistic development they just utter the name of the object that they have seen or heard, omitting the verbs: see or hear, the modal verb can and the subject I. For example: E (1;4) More hats Horns
‘I can see more hats’ ‘I can hear horns’
Apart from sentences aimed at making requests and drawing the listener’s attention to things or people, I have found few utterances in which content verbs are left out: E (1;2) Shoe Ha bath
‘I am putting on a shoe’ ‘Harry takes a bath’
E (1;4) Mamma
’Mamma sits here’
E (1;6) Peas Tea Kitty
‘She is shelling peas’ ‘She is making tea’ ‘She is feeding kitty’
Generally speaking, the omission of verbs which belong to the category of contentives is relatively infrequent. In my opinion the reason for this is semantic in nature. The verb is the central element of the sentence because it constitutes the essence of the message by providing information about what is happening to the subject or the action that the subject is performing. Because of this, a sentence deprived of its verb has hardly any meaning. For example, the utterance, J (2;0): Horsie baby, is completely unclear because the verb is missing. On the other hand, when other sentence elements are missing, they can be inferred from the context more easily on the basis of what the child was talking about before or from situational clues. If I consider, for example, the sentence J (3;0): Reading book the agent can be guessed from the context. In this particular situation it was he, because the boy was making up stories about his bear all the time; thus, the sentence can be easily expanded into: ‘He is reading a book’. This phenomenon has also been widely researched by many linguists adopting a cognitive approach to first language acquisition (Tomasello (1992), 69
Lieven, Pine and Baldwin (1997), Pine and Lieven (1993), Pine, Lieven and Rowland (1998)). They have formulated the Verb Island Hypothesis according to which the speech of young children is structured and organised around verbs. This theory proposes that the syntactic competence of young children is made up of the constructions specific for particular verbs. In other words, it is verbs that determine the form of the sentences. This means that if verbs are considered to be the focal point of the children’s language, they will be rarely deleted. However, all this should not lead us to the conclusion that, on the whole, verbs are hardly ever omitted in English sentences, since verbs belonging to the category of function words are omitted quite often.31 In my language corpus there are numerous examples of utterances deprived of such verbs, e.g. E (1;6) I shoes on I eat you This broken
‘I have the shoes on’ ‘I’ll eat you’, etc. ‘This is broken’
As I can see, the above sentences are not meaningless because even though the verbs are missing, they can be easily inferred from the context. What is more, the verbs omitted merely modulate the meaning of these utterances; therefore, what children say is quite understandable even if the verbs have been left out. Since children probably realise that the omission of these verbs causes no breakdown in communication, they do not feel obliged to use them in a sentence (Scholes (1970)). When it comes to the Polish language, verbs are omitted by children quite often although not as often as nouns functioning as agents or objects, which are deleted most frequently. The explanation of this phenomenon may be twofold. First of all, in inflectional Polish the grammatical case of the noun is marked by its inflectional endings and on the basis of that the meaning of the utterance may still be inferred, at least to some extent, even if the verb is missing. Moreover, in most cases, children’s words are accompanied by various situational clues, such as gestures, eye movement, etc., which facilitates the comprehension of their utterances. For example, the meaning of Z (1;8) Łyżeczkę ‘The spoon’ is less ambiguous than would be the same utterance of any English-speaking child since the inflectional fem. sg. ending -ę provides us with the information that this noun has been used in the accusative case and thus the missing verb is probably Daj or Chcę ‘Give me’ or ‘I want’. Another example would be Z (1;8) Serduszko ‘Heart’ standing for ‘Narysuj serduszko’ ‘Draw a heart’ said while the 31
This phenomenon is discussed in detail in the section devoted to the omission of various sentence elements in telegraphic speech in chapter 2.4.1.
70
girl was engaged in drawing, looking at her mother at the same time. Because the noun serduszko ‘heart’ has been used in the accusative case this utterance could be interpreted in the above context as a request. We can also offer yet another explanation of this phenomenon limited, however, to English utterances containing nouns derived from verbs by means of conversion, e.g. to kiss – a kiss, to walk – a walk, etc. In this case the verb which is present may only seem to be present. This is so because of the fact that the sentence: I kiss, on the surface, appears to contain a verb and mean: ‘I’ll kiss’, however, it may equally well be considered to contain a noun: ‘I’ll give you a kiss’ due to the existence of conversion in English. Thus, in a certain percentage of all children’s utterances the verb may only be thought to be present because of this misinterpretation. When it comes to the children in the Brown’s sample, only Eve showed an equal frequency of agent-action, action-object, and agent-object utterances; the other children, Adam and Sarah, did not produce agent-object utterances at all in Stage I. Bloom et al. (1975) also recorded agent-object sentences in only two of their four subjects. Other studies of the acquisition of English show that for some agent-action strings predominate (e.g. Kendall (Bowerman (1973)); whereas for others action-object predominate (e.g. Adam (Brown (1973)). My data confirms these findings in every respect, i.e. agent-object sentences are virtually non-existent, as I have found just two of them: one from Polish K (2;0) Nogą dzidziuś‘ [Baby with his leg] for ‘Dzidziuś rusza nogą’ ‘The baby is moving his leg’; the other from English: E (1;6) Ella for daddy for ‘Ella has made some tea for daddy’. With regard to the prevalence of agent-action or agent-object strings, it varies from child to child, e.g. Z. demonstrated a higher incidence of agent–action utterances, E. resorted to the action-object pattern more often, and K. relied equally frequently on both models. Therefore, there must be some other factors which account for such considerable divergences in the overall pattern of deletions. According to Maratsos (1984), it is individual differences and pragmatic factors that condition this process. 2.3.2.3. Individual differences and pragmatic factors
As regards individual differences, several researchers have tried to establish the existence of various styles and strategies of language acquisition that children resort to, which have a direct influence on the kind of sentence constituents used while omitting others. Bloom (1973) focuses on two approaches to the acquisition of syntax: pivotal and categorical. Children adopting a pivotal strategy use a small set of words responsible for expressing a constant functional relationship to the words they
71
combine with, e.g. more + X encodes recurrence, my +X possession, and this +X nomination. The pivotal strategy theory is in accordance with cognitive theories of first language acquisition according to which early child speech is organised around lexical elements (Tomasello (1992)). This means that children perceive the language not in terms of grammatical categories but in terms of lexical items responsible for the structure of utterances. According to Lieven et al. (1997) and Pine et al. (1998) many children make use of structures constructed around specific lexemes. My data confirms this, as it includes numerous utterances based on the use of particular words, e.g. no + noun, or verb + there. The words responsible for the organisation of a particular type of grammatical constructions are not used in other grammatical constructions. This suggests that they are not regarded as abstract grammatical entities but considered to be context-specific lexical items. In contrast, children who follow a categorical strategy generate sentences in which the grammatical or semantic relations between their constituents do not depend on the lexical items used. Thus, any noun, e.g. Mummy, may enter into all sorts of possible relations: Agent + Action, Possessor + Possessed or Action + Patient. The categorical strategy contradicts the main assumptions of cognitive theories of language acquisition, as, according to them, children below the age of three are not capable of forming abstract grammatical categories (Tomasello (1992)). On the other hand, the fact that the child is unable to form abstract grammatical categories does not mean that he is unable to combine words, as my data clearly proves the existence of some utterances actively constructed by children. Consequently, I believe that children make sentences without assigning grammatical categories to their constituents. The initial discrepancy presented above also remains as children begin to acquire more semantic relations. Bloom et al. (1975) has observed that children who initially follow the pivotal strategy, use pronouns and other proforms to express semantic roles although they know many of the names of the objects or people referred to and use those words in single utterances. However, in word combinations they use it, this, one or that to refer to patients, here and there for location, etc. When more semantic relations enter into their speech, they continue with the same approach, i.e. they use proforms combined with content words. Bloom (1975: 19) has concluded that the grammar the children are learning […] consists of relations between different verb forms and a number of constant functional forms. On the other hand, children following the categorical strategy use far fewer pronouns because semantic roles in their speech are filled by different nouns. This discrepancy continues until the MLU reaches about 2.5, when the utterances of the two groups of children begin to look the same. Bloom et al. (1975: 35) have proposed that […] children can break into the adult 72
linguistic code in one of (at least) two ways: with a system of formal markers, or with a system of rules for deriving grammatical categories. This means that children rely either on specific lexemes, or grammatical knowledge to organise their speech. Nelson (1975) has traced the differences in the strategies implemented by children in their approach to the language in the one-word stage. She classified children into two groups: referential and expressive. Referential children are those who in the initial stages of language acquisition spend a lot of time naming objects and in whose speech names for objects predominate. Consequently, they use a much higher proportion of nouns than pronouns in their multiword utterances. Expressive children seem to be more interested in social interaction and for that reason their early vocabulary is made up mainly of verbs and socialexpressive words, such as hi, more allgone, etc. As a result, their word combinations contain a more evenly balanced proportion of nouns and pronouns. Another possible source of differences in the pattern of deletions is thought to be of a pragmatic nature. It has been remarked that pragmatic factors have a considerable impact on the prevalence of certain semantic notions in two-word utterances and on the ordering of elements in those utterances. First of all, the context of the conversation influences the frequency of particular relations. For example, if the mother of the child concentrates in her interaction on eliciting names for objects, the child will frequently use forms expressing nomination (Brown (1973)): Jk (1;8) Cat Car Chair Dog
‘This is a cat’ ‘This is a car’ ‘That is a chair’ ‘That is a dog’
Children who are preoccupied with attracting their mothers’ attention will frequently dwell on notice, nomination and recurrence and those concerned with exploring the environment produce a great many utterances encoding locative action, state or attribution (Lieven (1978)). That was, in fact, the case with my daughter, who produced many sentences of the latter type: K (2;0) Szafa zepsuta [Wardrobe broken] Tam zapałki [There matches]
‘Szafa jest zepsuta’ ‘The wardrobe is broken’ (attribution) ‘Tam są zapałki’ ‘There are matches there’ (locative state) 73
Czerwone Sylwii [Red Sylvia’s] Duże buty [Big shoes]
‘Auto Sylwii jest czerwone’ ‘Sylvia’s car is red’ (attribution) ‘Tubisie mają duże buty’ ‘Teletubbies have big shoes’ (attribution)
K (2;1) Tutaj brzuszek mamy [Here mummy’s tummy] Krzesłem jestem [I am chair] Szybą jest [Pane is] Choinka tu [Christmas tree here] Rączkę tutaj [Hand here]
‘Tutaj jest brzuszek mamy’ ‘Here is mummy’s tummy’ (locative state) ‘Jestem pod krzesłem’ ‘I am under the chair’ (locative state) ‘Samochód jest za szybą’ ‘The car is behind the pane’ (locative state) ‘Tu jest choinka’ ‘Here is the Christmas tree’ (locative state) ‘Narysuj tutaj rączkę’ ‘Draw a hand here’ (locative action)
Constant interaction with siblings, especially those of a similar age leads to the creation of many sentences encoding possession (Schaerlaekens (1973)). Finally, the situational context may dramatically increase the frequency of otherwise rare, semantic notions. Retherford et al. (1981) notes that when his children were playing with play-dough they resorted quite often to notice or nomination. Although all these studies show a direct link between the pragmatic context and the frequency of semantic notions in conversation, they do not provide any data or evidence of whether these pragmatic factors influence the order of emergence of those notions in children’s speech. Bates and MacWhiney (1978) have investigated the relationship between the topic-comment function in language (Italian, Hungarian and English) and various syntactic devices, e.g. word order. It turned out that Italian children place comments (new information) in initial position, whereas English children place agents at the beginning of agent-action strings irrespective of their pragmatic role. In order to mark new information English children most often employ contrastive stress (Wieman (1976)). Polish children are often similar to Italian ones in this respect, e.g. K (2;3) Adzi to jest [Adzia’s this is] for ‘To jest dla Adzi’ ‘This is for Adzia’. Thus, they resort to a marked sentence pattern when they want to pass on new information about the subject because, according to Szwedek (1976), an unmarked sentence pattern is characterised by providing new information in the final position, whereas given information is in the initial position. Some other examples of employing a marked word order that I have 74
observed in the speech of my friends’ children are: Pluto pies ‘Pluto dog’ and Puchatek Kubuś [the pooh winnie] for ‘pies Pluto’ ‘The Pluto dog’ and ‘Kubuś Puchatek’ ‘Winnie the pooh’.32 To sum up, the communicative context in which the child’s early language is acquired as well as the cognitive and personality characteristics of both the child and his parents do affect the pattern of acquisition. The form of children’s utterances is determined by a few factors. First of all, it depends on the aims that the child wants to achieve by communicating with other people, such as attracting his mother’s attention, describing reality around him, emphasising what he considers important or underlining that a particular thing is his property. Secondly, the pattern of acquisition is affected by whether the child is socially oriented or not, i.e. whether he is interested in social interaction or not. Evidently, these findings go against generative theories of language acquisition, stating that children acquiring a language go through exactly the same stages at the same rate.
2.4. Ambiguity Ambiguity is an inherent feature of every natural language. If I think of ambiguity in the communication process, which is, by its nature, reciprocal, it is both sides that encounter this phenomenon. Thus, if I take into consideration interaction during the process of first language acquisition, it is both parents and children that need to disambiguate what they hear. However, when it comes to the early stages of development, which are of primary concern here, it is parents who face the task of disambiguating the sentences much more often since at this stage ambiguity results from the omissions of various morphemes and sentence elements by children. These sentences can be reconstructed more or less successfully thanks to the fact that language is characterised by redundancy.33 32
The issue of word order in the light of the rule of syntactic iconicity is discussed in chapter III. If I think of linguistic ambiguity in a broader sense, I usually have either grammatical or lexical ambiguity in mind (Dictionary of Language Teaching and Applied Linguistics (1999:15)). The following sentence is an example of grammatical ambiguity: He asked how old Sam was. Which can mean either: a. he wanted to find out the age of Sam. or b. he wanted to know whether old Sam was in good health. Lexical ambiguity results either from a. polysemy: She wears light trousers in summer. where light can mean either bright or thin. 33
75
Although the redundancy of natural languages exceeds 50% (Polański (1971)), which means that communication is possible with under half of the information of whatever sort being destroyed or just deleted, the recipients of incomplete sentences produced by children may have difficulties in understanding the message fully and thoroughly, the more so because in the initial stages of language development, the information deficiency is much higher than 50% or even 60%. Therefore, parents and caretakers are forced to make assumptions or guesses about the intended meaning of the utterances and, as a result, they come up with multiple interpretations. The number of possible interpretations is closely related to the degree of kinship or intimacy between the speaker (the child) and the listener. Obviously, the better the speaker and the listener know each other, the more restricted the range of meanings that is obtained since the utterances to be analysed are placed in a specific context. For this reason, the listener closely related to the child, in most cases, knows what the child wants to say and he also knows how to fill the information gap. However, an outsider interprets such utterances in many different ways because he is not familiar with the child’s way of expressing himself. In my further considerations I am going to analyse the most typical patterns of incomplete utterances from Polish and English data. I am going to study them independently of their context and check the degree of their ambiguity by trying to come up with possible interpretations. In my attempt to disambiguate Polish and English sentences I am going to provide as comprehensive an analysis as possible. At the same time, however, I am going to adopt the principle of conventionality,34 i.e. I will adhere to the conventional structures and forms that are used by the child at the two- and three-word stage. This means that I will not concentrate on all theoretically possible variants, as in or b. homonymy: Where is the band? However, this kind of ambiguity is beyond the scope of my considerations since it is encountered mainly in adult language not child language. Nonetheless, its significance has been recognised for first language acquisition, e.g. by Nerlich B. & D. D. Clarke (2002), who have studied the comprehension of polysemous utterances by children and argued that this ability comes as one of the last stages of linguistic and cognitive development. 34 Nerlich (2002: 66) has stated that […] in linguistics and psycholinguistics it has generally been assumed that I adhere to the principle of conventionality, that is, that I stick to the conventional meanings of words. Although the principle of conventionality is discussed there from the semantic perspective, I am going to adopt it for my grammatical analysis because the nature of the disambiguation is the same in both cases; namely it is disambiguation in context, i.e. interpreting words or whole utterances in context. This means that in the case of telegraphic speech it is disambiguation of the utterances in the context of the child’s grammatical competence.
76
many cases their number would be endless, but only on those that are psycholinguistically plausible for early child language. In my analysis I will first deal with the ambiguity of the sentences deprived of function words and then those that lack content words. 2.4.1. Ambiguity of telegraphic speech First of all, it might seem that telegraphic speech is characterised by lesser ambiguity than utterances deprived of content words because in the case of telegraphic speech deletion affects grammatical morphemes, which are limited in their number, and therefore the number of possible interpretations is also limited. Even more limited than the number of grammatical morphemes is the child’s knowledge of grammatical constructions. This is the key issue, in fact, because if sticking to the principle of conventionality in the disambiguation process only those grammatical structures that are psycholinguistically plausible (i.e. ones that the child is familiar with and consequently can use) can be relied on. Thus, a single word, such as, for example, play could theoretically be expanded into numerous sentences, finite or non-finite, by adding all functors that would form grammatically correct utterances. For example, it could be expanded into any tense: I have been playing, I will be playing, I had played, etc. or any other form, for instance perfect infinitive: I could have played, I needn’t have played, etc. or participle: playing, having played. However, practically speaking, this sentence has just a few expansions because at this age there are just a few grammatical constructions that the child knows.35 Therefore, the intended meaning of this sentence could be the following: I am playing, I will play, I played, Can I play? etc. Evidently, words used by children denote quite broad grammatical categories, as the verb play stands for ‘the action of playing’ unspecified as to its tense, agent, aspect, mood, etc. Consequently, the infinitive play can be regarded as the prototypical, i.e. basic element, serving as a basis for the formation of more sophisticated constructions. When taking a look at some other telegraphic utterances produced by children, it turns out that the elements used in them are prototypical as well. For example, in the English sample the nominative noun functions as a basic category and it may stand for plural number, prepositional phrase, declarative sentence, question, etc. The presence of prototypical elements in the grammatical development has been discovered by cognitive linguistics (Taylor (1989), Meints (1999)), however, no detailed research has been carried out in this field so far. Nonetheless, some child language researchers have proposed that language acquisition is based on the so called radial strategy 35
The order of the acquisition of grammatical morphemes will be discussed in chapter III.
77
(Sinha (1994)).36 This means that children first acquire the cognitively and grammatically simplest forms only to broaden their repertoire with more complex and sophisticated forms. On the basis of my data it is evident that the radial strategy is employed by Polish and English-speaking children in the period of telegraphic speech. Telegraphic speech utterances abound in prototypical forms and constructions to be developed into more complex ones in further stages of linguistic development. Thus, the period of telegraphic speech can be regarded as the prototype-oriented stage of language acquisition, as it contains only very basic grammatical forms. Consequently, the simplicity of grammatical forms leads to their considerable ambiguity. Another characteristic feature of telegraphic speech that is tantamount to its ambiguity is its high degree of schematicity (Langacker (1987)). This means that the level of specificity, i.e. precision, is quite low. Langacker accounts for this fact, claiming that the wider scope, the lower level of specificity. Langacker states that scope should be understood as the conceptual proximity to or distance from specific linguistic items. Thus, the greater the linguistic and cognitive distance from particular linguistic items, the greater the schematicity of language. When talking about a significant distance from a particular word or construction in the process of first language acquisition, a limited mastery of the child’s native language and the rules governing its structure is meant. This, in turn, leads to a limited use of language, i.e. its schematicity. As my data shows, the language of children is very schematic, as it is capable of expressing only a restricted number of concepts and messages that can be interpreted in a wide variety of ways. As far as this analysis is concerned, I am going to delineate the most common patterns of telegraphic speech and concentrate only on those utterances which are potentially ambiguous, disregarding all those with no scope for ambiguity. First, I will present the major types of ambiguous telegraphic speech from Polish and English data. Next, I will compare the omissions and their expansions,37 adopting for that purpose a grammatical classification, believing that it will reflect best both the difference between the morphological systems of the two languages and the kind of structures acquired by children in the early stages of their linguistic development. In addition, I am also going to dwell on the meaning of telegraphic utterances 36 Sinha has discovered reliance on the radial strategy in the process of the acquisition of English locative particles. However, my data clearly indicates that this strategy is made use of on much a wider scale. The notion of radial strategy has been derived form the concept of radial category proposed by Lakoff (1987). 37 Expansion is a term first used by Brown and Belugi (1964), who meant by it adult interpretations of children’s sentences, i.e. what in view of all the circumstances, the child meant and ought to have said.
78
and check whether semantic notions that are acquired at this stage are language specific or cross-linguistically universal. 2.4.1.1. The most common patterns of ambiguous telegraphic speech from Polish and their expansions
The order of the utterances presented in this section has been established on the basis of their frequency in my sample, beginning with the most common patterns. Patterns I and II are prevalent in the one- and two-word stages: pattern I can be found in about fifty per cent of all the utterances and pattern II constitutes about twenty per cent of the whole sample. The frequency of these two patterns decreases as the children enter the three-word stage when their speech is made up of other elements as well, such as agent, action, object, locative phrase, etc. Patterns III and IV are quite common in the two- and threeword stage, constituting about fifteen percent of all the utterances. Pattern V and VI can be found mainly in the two-word stage and their frequency is relatively low, about ten per cent. Lastly, Pattern VII and VIII are quite infrequent, representing only five per cent of all the utterances. The examples presented here come from all the three children studied and they were chosen if, for the sake of their universality, the following two criteria were met: 1. They were found in all the three children. 2. They occurred systematically in the sample, even if their frequency was quite low. This section will be devoted to presenting and then clearing up the ambiguity of the telegraphic utterances from Polish. With a view to explaining their meaning I am going to provide two types of expansions: a and b Expansions a stand for what was really meant by a particular utterance. Expansions b represent theoretically possible meanings of an utterance inferred by means of the interpretations of the utterances of the same type by the same child (in different circumstances) or by other children. These are the most common patterns of ambiguous telegraphic speech from Polish and their expansions: I. Nominative nouns: Osioł, piłka, piosenka a. ‘To jest osioł’, etc. b. 1. ‘Czy to jest osioł?’ 38 2. ‘To jest mój osioł’.
‘A donkey, a ball, a song’ ‘This is a donkey’, etc. ‘Is this a donkey?’ ‘This is my donkey’
38
Practically speaking, any of the utterances presented here can theoretically be expanded into a question, as in Polish it is possible to convert a declarative sentence into a question by means of rising intonation, without changing the word order. However, only some utterances were intended as questions although they did not have a question-like structure and only they have been incorporated into the analysis.
79
II. Adjectives: Gorące, zimne, ciężkie a. ‘To jest gorące’, etc. b. 1. ‘Czy to jest gorące?’ 2. ‘One są gorące/To jest gorące’, etc.
‘Hot, cold, heavy’ ‘This is hot, etc.’ ‘Is this hot?’ ‘They are/it is,etc. hot’
III. Infinitive (imperfective) + adverbs: tutaj, tam ‘here’, ‘there’: Malować tutaj ‘Paint here’ a. ‘Będę tutaj malować’ ‘I will paint here’ b. 1. ‘Trzeba to tutaj pomalować’ ‘This should be painted’ 2. ‘Muszę tutaj coś namalować’ ‘I must paint something here’ 3. ‘Czy mogę tutaj malować?’ ‘Can I paint here?’ 4. ‘Masz tutaj coś namalować’ ‘You are to paint something here’. With regard to the use of infinitives in early child speech it has been remarked that it is caused by the use of numerous impersonal verb forms in the language addressed to them, e.g. Będziemy malować ‘We will paint’, Nie możesz malować ‘You cannot paint’, etc. (Wexler (1994), Leonard (1998)). Thus, children master either the whole infinitive constructions or regard the infinitive as the prototypical verb form and overgeneralise its use. If the last utterance from the above set, i.e. Masz tutaj coś namalować ‘You are to paint something here’ is to be considered, an alternative expansion could be: ‘Mama ma tutaj coś namalować’, i.e. ‘Mummy is to paint something here’. The reason for this is that in the early stages the child explicitly marks the agent in the sentence by expressing it with nouns, common or proper, rather than pronouns, e.g. Kinga nie umie ‘Kinga can’t’ or Mama umie ‘Mummy can’ instead of Ja nie umiem ‘I can’t’, Ty umiesz ‘You can’. However, since with the passage of time, the other pattern becomes more and more common and finally replaces constructions with the explicit agent, I have decided to stick to it. Another thing that needs to be mentioned at this point is that malować ‘paint’ in the sentence Malować tutaj ‘Paint here’ is an example of the imperfective verb. In the case of imperfective verbs, devoid of prefixes, and meant to be used by the child in a perfective sense, it is much more difficult to clear up the ambiguity of the sentence since one of the missing word-forming prefixes has to be provided, which, apart from being the markers of perfectivity, give the verb a new lexical meaning, e.g. na-malować, po-malować, zamalować, etc. Thus, the ambiguity of the utterances containing imperfective verbs is both of grammatical and derivational nature.
80
IV. Infinitive (perfective) + direct object: Światło zaświecić [The light to turn on] a. ‘Trzeba zaświecić światło’ ‘The light should be turned on’ b. 1. ‘Czy mogę zaświecić światło?’39 ‘Can I turn the light on?’ 2. ‘Masz zaświecić światło’ ‘You are to turn the light on’ In the utterance Światło zaświecić [The light to turn on] a marked word order is used, i.e. object-verb instead of the unmarked one, i.e. verb-object: Zaświecić światło ‘To turn the light on’. As has been remarked in section 2.3.2.3, Polish children often employ a contrastive word order so as to emphasise the new or the most important information, i.e. they shift the words denoting new or essential information to the beginning of the sentence. In this particular case the object światło ‘the light’ has been moved from its customary position in order to draw the listener’s attention to the fact that it is the light, not something else, that needs to be turned on. 40 V. Incomplete prepositional phrase: I have decided to refer to this pattern as the incomplete prepositional phrase because the original children’s utterances meant as prepositional phrases are most often devoid of prepositions, and in this sense they are incomplete. Prepositions are inferred from the context and present only in the expansions. Another thing is that the very classification of this pattern is ambiguous in itself since, as it appears from my data, the same utterance may be categorised as both an incomplete prepositional phrase and a noun in the genitive case: babc-i ‘granma’s’ for ‘To jest babci’, i.e. ‘This is granma’s’. However, if such utterances are treated as nouns in the genitive case, they are not ambiguous and, because of this, beyond the scope of my considerations. Therefore, I have decided to label this pattern an incomplete prepositional phrase. Babci a. ‘Do babci’ b. 1. ‘U babci’ 2. ‘Dla babci’ 3. ‘To jest babci’ 4. ‘Byłam u babci’
‘Granma’ ‘To granma’ ‘At granma’s’ ‘For granma’ ‘This is granma’s’ ‘I was at granma’s’
39
When it comes to the interrogative expansions of this utterance, the polite request Can you......? is a form that the children are not familiar with at this stage. It has therefore been excluded as a potential interpretation here. 40 A discussion of the significance of word order can be found in chapter III in the section devoted to syntactic iconicity.
81
Incomplete prepositional phrases, like those above, are very often answers to questions, e.g. Gdzie jutro idziemy? Babci for ‘Do babci’, i.e. ‘Where are we going tomorrow?’ ‘Granma’ for ‘To granma’, or Dla kogo to jest? Babci for ‘Dla babci’, i.e. ‘Who is it for?’ ‘Granma’ for ‘For granma’. However, the same telegraphic utterance may, in other circumstances, stand for the whole sentence: Chcę iść do babci or Chodźmy do babci, etc., i.e. ‘I want to go to granma’ or ‘Let us go to granma’, etc. Thus, the same word or a string of words can be both an example of telegraphic speech and an utterance devoid of content words. As far as the prepositions used in the above expansions are concerned, their selection is conditioned by the case of the noun in the incomplete prepositional phrase. In Polish prepositions govern noun cases, e.g. u ‘at’, do ‘to’, od ‘from’, dla ‘for’, bez ‘without’ require the genitive, which has the widest range of functions among Polish cases. Therefore, the choice of prepositions in Polish expansions of this type is restricted by the noun used in the particular case by the child. VI. Accusative noun: Koronę a. ‘Ja mam koronę’ b. 1. ‘Daj mi koronę’ 2. ‘Ja chcę koronę’ 3. ‘Narysuj mi koronę’
‘A crown (acc.)’ ‘I have a crown’ ‘Give me the crown’ ‘I want a/the crown’ ‘Draw a crown for me’
This pattern can be treated as either an example of telegraphic speech or a sentence devoid of content words. Thus, because these two kinds of omissions very often overlap I have decided to incorporate here the examples of sentences with the content words deleted. Another reason for this is their very high frequency in the sample. Moreover, I believe this gives a broader view of the ambiguity of this particular utterance. VII. Subject +imperfective verb (+ object): Zuzia piła to ‘Zuzia was drinking it’ a. ‘Zuzia będzie piła to’ ‘Zuzia will drink it’ b. ‘Zuzia wypiła to’ ‘Zuzia drank it’ This particular pattern is an example of the utterance containing the socalled light verb, i.e. the verb which is frequently used and which denotes a very general meaning. Light verbs play an essential role in the process of first language acquisition as they are very frequent in the language input because of their semantic generality. The more general the semantic meaning the verb 82
expresses, the more frequently it can be used by adults and imitated by children. What is more, light verbs encode meanings that are particularly important in our everyday life: they represent patterns of action and movement, relations of cause and effect, and transfer. These basic patterns and relations can be regarded as building blocks to be used to express the majority of human cognitive experience in the process of abstraction and metaphorical extension (Lakoff (1987), Goldberg (1995)). Thus, the use of light verbs during the period of telegraphic speech constitutes a necessary prerequisite for further cognitive and linguistic development. Therefore, it is light verbs that are first acquired in many languages of the world, such as Finnish (Bowerman (1973)), French (Gregoire (1937)), Japanese (Sanches (1978)), and Korean (Park (1977)). My data clearly shows that in the period of telegraphic speech both Polish and English-speaking children rely heavily on light verbs that are both semantically and cognitively prototypical. As far as the construction of the above pattern is concerned, it can again be stated that it possesses a dual structure. First of all, it can be interpreted as a complete and well-formed utterance referring to a past action of some duration: Zuzia piła to ‘Zuzia was drinking it’. Secondly, it may, as it is here, be an example of telegraphic speech. VIII. Reflexive finite verb without the reflexive pronoun: Uderzył ‘He hit’ a. ‘Uderzył się’ ‘He hit himself’ b. ‘Uderzył mnie, jego’, etc.41 ‘He hit me, him, etc.’ Having presented the most frequent patterns of telegraphic speech form Polish, let me now pass on to my English sample. 2.4.1.2. The most common patterns of ambiguous telegraphic speech from English and their expansions
As in the case of the patterns of ambiguous telegraphic speech from Polish, I have ordered the utterances presented below according to their frequency in the sample. Pattern I and II are prevalent in the one- and twoword stages, constituting approximately half of all the utterances. When the children become more advanced in their linguistic development, pattern I disappears from their speech and the frequency of pattern II gradually decreases. Pattern III and IV are typical for a two- and three-word stage. Pattern III accounts for approximately twenty per cent of all the utterances, 41
Obviously, the choice of the appropriate object pronoun depends on the context, i.e. the person affected by the action.
83
whereas pattern IV can be found in about ten per cent of all the sentences produced by children. Pattern V and VI are relatively rare with the frequency amounting to about five per cent. The criteria adopted for choosing particular patterns and providing their expansions are the same as in the case of the analysis of the patterns of ambiguous telegraphic speech from Polish. These are the most common patterns of ambiguous telegraphic speech from English and their expansions: I. Nouns 1. Animate nouns: Doggie a. ‘This is a doggie.’ b. 1. ‘This is my doggie.’ 2. ‘Is this a doggie?’ 3. ‘These are doggies.’ 4. ‘For doggie, with doggie’, etc. 5. ‘(This is) doggie’s’ Apart from the expansions presented here for this particular case, I should remark that a singular noun can also denote a root compound (Clark 1993). An example of this phenomenon from my data would be J.(3;0) apple for ‘an applepie’. However, because examples of this kind are not representative of the whole sample, and relatively infrequent, I have decided not to include them in my expansions. 2. Locative nouns: Car a. ‘This is a car’ b. 1. ‘This is my car.’ 2. ‘Is this a car?’ 3. ‘These are cars.’ 4. ‘In the car, into the car, by car’, etc. At this point I would like to explain the origin and the meaning of the term locative noun which I have decided to adopt for the present analysis. The locative case of the noun in case grammar (Fillmore (1968)) refers to the location of the action expressed by the verb. Here I have decided to extend this term to the locative noun, i.e. a noun preceded by prepositions and describing the position of another noun or the direction of its movement. 84
The difference between the Polish and English classification of this pattern (in the Polish sample nouns of this type are classified as incomplete prepositional phrases) results form the fact that in Polish the inflectional morphemes, marking particular cases, tell us whether the noun can be a part of the prepositional phrase or not, whereas in English there is no such indication. For example, in Polish only the locative case: samochodzie ‘car’ suggests that it is an incomplete prepositional phrase meant to denote location. In English, lacking inflectional endings, the locative function can be assigned to practically any noun denoting some area or place. In such a case this noun is an example of an incomplete prepositional phrase, e.g. car for ‘in the car’. Apart from that, one-word utterances containing this type of noun can also stand for a question: ‘Is this a car?’ or a declarative sentence stating either the existence of a given object: ‘This is a car’ or denoting possession: ‘This is my car’, etc. (see the expansions above). The next construction to be discussed at this point is an extremely common and ambiguous pattern, i.e. verb + object. II. Transitive/intransitive verb (+ dir. object): Transitive verb + direct object Intransitive verb Eat them42 Jump/sit a. ‘Can I eat them?’ a. ‘Can I jump/sit?’ b. 1. ‘I will eat them’ b. 1. ‘I will jump/sit’ 2. ‘I am eating them’ 2. ‘I am jumping/sitting’ 3. ‘Are you eating them?’ 3. ‘Are you jumping/sitting?’ 4. ‘I am going to eat them.’ 4. ‘I am going to jump/sit’ 5. ‘Let’s eat them’. 5. X 6. ‘Did you eat them? 6. X 7. ‘You should/are to eat them.’ 7. ‘You should/are to jump/sit’ 8. X 8. ‘I jumped’/ X 9. ‘He/she eats them’ 9. ‘He/she jumps/sits’ Altogether there are four possible combinations in this pattern: regular verb+ object, irregular verb+object, regular verb and irregular verb. In this sample the three models presented above, i.e. irregular verb+object and regular/irregular verb are the most representative for the whole category because they are the most frequent. The reason for the high frequency of the irregular verb+object pattern is that the vast majority of transitive verbs mastered and used by the children in this sample in the period of telegraphic speech are 42
This utterance, besides being an example of telegraphic speech can also be a perceived as a grammatically correct sentence, namely the imperative.
85
irregular verbs, describing their everyday activities, such as drink, eat, go, find, take, see, hear, give, put, draw, break, make, etc. The regular verb+object pattern, e.g. E (1;11): Rock a my baby can also be found in my corpus but in the initial stages of linguistic development it is relatively rare. When it comes to intransitive verbs, I have found that the rate of acquiring intransitive regular verbs and intransitive irregular verbs is the same. As can be seen above, whether the verb is regular or not as well as whether it is followed by an object or not has a direct influence on the pattern of expansions. If a particular expansion is non-existent for a given pattern, I have decided to mark this with X. In Polish the above pattern is not ambiguous as the verb inflections are marked for gender, number and person and the ellipsis of the subject is allowed if it can be inferred from the context. The only ambiguity that may arise concerns the use of the third person singular, e.g. Wylała mleko ‘Poured out the milk’ because young children often use this form for self-reference, therefore, it may not be known whether they are talking about themselves or about somebody else, e.g. their parent. Thus, such a sentence can be interpreted as either: ‘Kinga wylała mleko’, i.e. ‘Kinga poured out the milk’ or ‘Mama wylała mleko’, i.e. ‘Mum poured out the milk’. Besides, in English, the same kind of omissions may be observed in this pattern when some other sentence elements are present as well, e.g. adverbs of place, time and manner: here, now, high, etc. (Eat here, Jump high), or prepositional phrases, e.g. from the plate, in the kitchen, etc. (Eat from the plate, Jump in the kitchen). As far as the cognitive structure of this pattern is concerned, it can be labelled as the billiard ball model (Langacker (1997)). The billiard ball model characterises the nature of actions with regard to the flow of energy. The basic assumption of the billiard ball model is that the source of energy is primal in relation to the event, i.e. the agens is primal in relation to the patiens. Consequently, this model finds its reflection in the structure of many utterances in which the cardinal role of the agens, understood as the source of energy, i.e. the triggering element, is marked by its initial position in the sentence. In contrast, the patiens is regarded as the recipient of energy, therefore it follows the agens and the verb that denote the transmission of energy. The reason for this is that the process of reception always follows the process of transmission. In the case of telegraphic speech utterances it is the verb denoting an activity that occupies the initial position in the sentence due to the fact that the subject is frequently omitted. The next two patterns I am going to discuss below are also constructed around verbs. However, they are far less ambiguous. 86
III. Present participle (+ object): Sleeping Eating it up a. ‘She43 is sleeping’ a. ‘She is eating it up’. b. 1. ‘I was sleeping’ b. 1. ‘I was eating it up’. 2. ‘Are you sleeping?’ 2. ‘Are you eating it up?’ IV. Noun + verb: Daddy sleep a. ‘Daddy is sleeping’ b. 1. ‘Daddy, I am sleeping’ 2. ‘Daddy, are you sleeping?’ 3. ‘Daddy, sleep’ The ambiguity of pattern IV lies not only in its vague meaning but also in its equivocal classification because there are cases in which it can be categorised either as noun+noun or noun+verb combination. This problem pertains to all strings composed of nouns+nouns derived from verbs by means of conversion, which is quite productive in English, meaning the number of such pairs is also significant. I would like to quote two examples from my sample. For instance, J (2;0) Daddy work once meant ‘Daddy is at work’, so it was a noun+noun combination but in some other situation it was used in a sense ‘Daddy is working’, i.e. noun+verb combination. Another example would be E (1;4) Rosie walk, which in my sample was meant to express both genitive: ‘Rosie’s walk’, referring to the title of the girl’s book and a present progressive action ‘Rosie is walking’. The remaining two patterns to be presented are the least frequent but not the least ambiguous. V. Noun + noun: Bear hat a. ‘That’s a bear’s hat’ b. 1. ‘There is a bear in the hat’ 2. ‘That’s a bear hat’44 3. ‘That’s bear and a hat’ 4. ‘That’s a hat for a bear’ 5. ‘The bear has got a hat’ 43
Again, the agent is rather ambiguous here since any other personal pronoun may be used, depending on the context. 44 This is the so called classificatory semantic relation (Brown (1970:179)), which consists of two nouns where the first noun is used in the function of an adjective, describing the qualities of the second noun. Here the bear hat refers to a hat which is, i.e. looks like, a bear.
87
VI. Adjective/ adjective + noun Crazy Big monster a. ‘I am crazy’ a. ‘I am a big monster’ b. 1. ‘I am being crazy’45 b. 1. ‘Be a big monster’ 2. ‘Is she crazy?’ 2. ‘Are you a big monster?’ All in all, there are considerable differences and striking similarities between Polish and English patterns of telegraphic speech both with regard to their structure and possible interpretations. The next section is a discussion of Polish and English telegraphic speech. 2.4.1.3. Grammatical analysis of Polish and English telegraphic speech
The above data has shown conclusively that there are differences both with regard to the most common patterns of telegraphic utterances and the degree of the ambiguity of telegraphic speech in Polish and in English. On the other hand, meanings expressed by children in the two languages remain basically the same.46 When it comes to the grammatical forms used in telegraphic speech and presented above, there are two basic differences between English and Polish. First of all, the number of grammatical morphemes is considerably smaller in English than in Polish due to the fact that English has an impoverished morphology in comparison with other languages 47 and, as a result, one form stands for many grammatical categories, whereas in Polish different verb or noun forms are marked with different inflectional endings. 48 Moreover, in English it is often enough just to add the inflectional morpheme to the basic form of the noun or the verb to change its grammatical category, e.g. the 45 46
Utterances of the b.1. kind have been found very rarely. For a discussion of the acquisition of semantic categories see the next subchapter, i.e.
2.4.2.4. 47
In the light of cognitive theories of language acquisition English exhibits a lesser degree of indexicality than Polish. By indexicality I mean the presence of indices in language, and these are, among other things, grammatical morphemes (Tabakowska (2001)). The issue of indexicality is discussed in detail in chapter IV. 48 In spite of the fact that Polish possesses a rich morphological system, which means that each grammatical form is marked by a particular morpheme, different from all the others, one inflectional morpheme may also occasionally represent two grammatical forms. For instance, non-virile, nominative plural nouns have the same suffix -i or -y as non-virile accusative plural nouns and as a result can function in a sentence either as agents or as objects: Matki czeszą dziewczynki ‘Mothers comb the girls’ hair’. If I regard this sentence as one with unmarked word order Matki ‘Mothers’ is the subject of the sentence but if I assume that the word order has been changed for pragmatic purposes Matki ‘Mothers’ then functions as the object, which has been emphasised by its being shifted to the front position. A detailed discussion of Polish conjugational and declensional forms can be found, for example, in Bąk (1989).
88
singular noun girl is changed into plural by adding the plural morpheme -s. Therefore, I can talk about the omission of some morphemes. In contrast, in Polish the basic form of the word, i.e. the bare stem is an abstraction because every vocabulary item is marked: nouns for gender and number, verbs for person, tense and also number. Thus, in Polish changing the grammatical category of the verb is brought about by replacing some bound morphemes by others, not by adding them. For example, when changing the singular noun: dziewczynk-a ‘girl’ into the plural: dziewczynk-i ‘girls’, the feminine, singular, nominative morpheme -a is replaced by the feminine, plural, nominative morpheme -i. Therefore, in the early period of child speech a failure to use certain morphemes by English-speaking children and the use of the wrong ones by Polish children can be observed. This is why English is a more telegraphic language than Polish: in English unfamiliarity with a certain grammatical category results in the omission of a grammatical morpheme, i.e. producing a telegraphic utterance, whereas in Polish it very often means using the incorrect affix. On the other hand, one can talk about using wrong inflectional endings in the telegraphic speech in Polish only in the case of bound morphemes, which constitute the vast majority of Polish grammatical morphemes. When taking free morphemes, such as the auxiliaries będzie ‘will’, and ma ‘has’, or personal pronouns ja ‘I’, ty ‘you’, etc., their omission also entails generating telegraphic speech utterances similar to those produced by English-speaking children, e.g. K (2;0) Misia ‘Teddy bear’ for ‘Mam misia’ ‘I have a teddy-bear’. However, since free morphemes are not numerous, the above phenomenon is not very common. The second difference is a direct consequence of the first one. If, in English, one form may express various grammatical relations, the meaning of such an utterance will be quite ambiguous – much more ambiguous than in the case in which every grammatical form possesses a specific inflectional ending. For example, let us take nouns, which in English have no other inflections than plural -s and genitive ‘s, whereas in Polish they are marked for gender, case (where genitive is just one of many possibilities) and obviously number, marked by different endings depending on the declension. Therefore, the expansion of an English telegraphic utterance made up of a noun is much richer than that of a Polish one. For example, in English the noun pattern: donkey can not only perform the naming function,49 express possessive relation or make a question but can also stand for a prepositional phrase or plural form. In Polish all these 49 According to Brown (1970) [...] the simple naming function is one of the most reliable in Stage I speech. This finding is strongly confirmed by my data, where all the six children at the onset of speech extensively named things and people.
89
grammatical forms cannot be expressed by the same pattern, because apart from adding inflectional endings they also involve making some changes to the noun stem, e.g. osł-y ‘donkeys’, osł-a ‘donkey (acc.)’ for ‘I have a donkey’ or na oś-le ‘on the donkey’. Therefore, they must be classified as different categories. With regard to the noun+noun combination, it can be found in English telegraphic utterances because of its universality, stemming from the fact that irrespective of the grammatical and semantic relation between the two elements, apart from the genitive, the nouns have the same form, i.e. the zero ending. In my Polish sample I have found no strings consisting of two nouns that would be examples of ambiguous telegraphic speech. There are noun+noun combinations but they either express an unequivocal possessive relationship – Z (1;11) Rowerek Krzysia ‘Krzyś’s bike’ for ‘To jest rowerek Krzysia’ ‘This is Krzyś’s bike’ – or are devoid not of function but of content words, e.g. K (1;8) Dzidziuś mamą ‘Baby mummy (instr.)’ for ‘Tam jedzie dzidziuś z mamą’ ‘There goes the baby with his mummy’. Sentences of the latter kind are ambiguous but they are not examples of telegraphic speech. The same observations, concerning the Polish and English morphological systems, apply also to patterns containing verbs, which, again, in English have very few inflections. Because of this, in the examples of telegraphic speech from English fewer patterns containing verbs can be differentiated than in Polish. However, these patterns offer more possibilities for their interpretation than the Polish ones. Thus, pattern II for English, i.e. verb (+object), Eat them, has nine expansions, i.e. it is ambiguous to a great extent because it is possible to add a wide variety of morphemes, either bound or free, to the verb stem, which leads to numerous grammatical constructions, e.g. ‘Can I eat them?’, ‘I will eat them’, ‘I am eating them’, ‘Are you eating them?’, ‘Did you eat them?’, etc. As far as Polish is concerned, I have distinguished three patterns of utterances containing verbs: pattern III, i.e., infinitive (imperfective) + adverbs tutaj, tam ‘here, there’, e.g. Malować tutaj ‘Paint here’, pattern IV, i.e. infinitive (perfective) + direct object, e.g. Zaświecić światło ‘To turn the light on’ and pattern VII, i.e. subject + imperfective verb (+object), e.g. Zuzia piła to ‘Zuzia was drinking it’. All these three patterns perform the majority of communicative functions from pattern II in the English sample:
Request for permission Order
90
Polish Expansions of patterns: IV, III and VII ‘Czy mogę tutaj malowac?’ (III) ‘Can I paint here?’ ‘Masz tutaj coś namalować’ (III) ‘You are to paint something here’
English Expansions of pattern II ‘Can I eat them?’ ‘You are to eat them’
Future action ‘Zuzia będzie piła to’ (VII) ‘Zuzia will drink it’ Past action ‘Zuzia wypiła to’ (VII) ‘Zuzia drank it’
‘I will jump’ ‘I jumped’
The communicative functions absent from the Polish expansions of these three patterns are those whose formation would involve changing suffixes added to the verb stem and changing the verb stem itself. As I have already stated, Polish is less telegraphic than English because inflectional suffixes can never be omitted, as they must always be present even in the basic form of nouns, verbs or adjectives. For example, in the telegraphic utterance: Światło zaświecić ‘To turn the light on’ the verb zaświecić ‘turn on’ ends in an inflectional suffix -ić, which in Polish marks the infinitive. Changing the grammatical category of this verb entails replacing the inflectional infinitive ending -ić with another one, e.g. Zaświeć-my światło ‘Let us turn the light on’ or Zaświeci-łaś światło? ‘Have you turned the light on?’ Therefore, expanding telegraphic utterances just by adding inflectional suffixes is not possible in Polish, which limits significantly the number of possible interpretations in each case. Below I have listed the expansions of English pattern II which have no their equivalents in the expansions of any of the three Polish patterns: III, IV and VII:
Present action Suggestion Question: a. present b. past
Polish Expansions of pattern II X X X X
English ‘I am eating them’ ‘Let’s eat them’ ‘Are you eating them?’ ‘Did you eat them?’
To sum up, each of the patterns from the Polish sample, III with four expansions, and IV and VII with two expansions, is far less ambiguous that the whole English pattern II which has as many as nine expansions. As far as the classification of the telegraphic speech patterns containing verbs is concerned, the differences between Polish and English in this respect result from the way of forming perfective and imperfective verb forms in the two languages. The Polish verb is either perfective or imperfective with the imperfective form being morphologically simpler. The perfective form is, in the majority of cases, derived from it. When it comes to the way in which aspectual differences are realised, it can be either through prefixation, e.g. pisać/na-pisać ‘to write’ (imperf./perf.), suffixation, e.g. kop-a-ć/kop-ną-ć ‘to kick’ 91
(imperf./perf.) or suppletion, e.g. brać/wziąć ‘to take’ (imperf./perf.) (Majewicz (1982)). In English the majority of verbs are neutral with respect to aspectual meaning and only using them in, for example, the perfect tense makes them perfective: I have read, I had read, I will have read, I could have read, etc. (Fisiak (1978)). Thus, unlike in Polish, the very addition of the auxiliary have can change the verb form into the perfective one. Another issue that has influenced my classification is the way of forming the future tense with perfective and imperfective verbs in Polish. For perfective verbs forms corresponding to present imperfective verbs are used, e.g. na-piszę ‘I will write’ (perf.) – piszę ‘I write’(imperf.), po-jadę ‘I will go (perf.) – jadę ‘I go’(imperf.), where the suffix -ę is the first person singular ending for the two forms. However, with imperfective verbs an analytic construction is used, which consists of the future form of the auxiliary być ‘to be’: będę ‘I: will’, będzie ‘he/she: will’, będziemy ‘we: will’, etc. followed by either the infinitive, będę malować ‘I will paint’, or the past verb form, będę malowała ‘I will paint (past, 3rd SG)’, będziemy malowali ‘We will paint (past, 3rd PL)’, etc. Therefore, two kinds of telegraphic utterances with the auxiliary będzie omitted, i.e. referring to future, have been classified separately because they are made of different verb forms: Zuzia piła to ‘Zuzia was drinking it’ (pattern VII) 50 or Zuzia pić to ‘Zuzia drink it’ and, consequently, they can be expanded in two different ways. In English there is only one form of the future tense, formed by putting an auxiliary in front of the infinitive. Lastly, as far as adjectives are concerned, in Polish they are marked for gender and number so it is obvious, or at least it may be inferred, to whom or what they refer, whereas in English the gender and the number of the referent is ambiguous. Therefore, this pattern is more ambiguous in English than it is in Polish. To sum up, English telegraphic speech is much more ambiguous than Polish51 because of the fact that in English there is no one-to-one correspondence between the form and the meaning and one form may express a variety of grammatical concepts.52 This, in turn, leads to the inconclusive grammatical and semantic interpretation, as my analysis of the telegraphic utterances has shown. 50 Smoczyńska (1985) has found no auxiliary deletion with the past participle. I have found such forms in my sample, but very infrequently; therefore, the frequency of pattern VII, in which they can be found, is low. 51 For a discussion of ambiguity in adult Polish and English see the subchapter devoted to the analysis of block language on the basis of data from the two languages. 52 In the light of the theory of Natural Morphology presented in chapter III this phenomenon is referred to as lack of uniformity and transparency of encoding.
92
The next section is devoted to a semantic study of telegraphic speech with a view to comparing the semantic categories expressed by children in Polish and English and the order of their acquisition. 2.4.1.4. Semantic analysis of Polish and English telegraphic speech
The reason for carrying out a semantic analysis of telegraphic speech is the fact that according to cognitive linguistics early child speech is organised almost exclusively around lexical elements. This means that children’s speech is semantically-oriented, i.e. the child’s attention is concentrated on specific lexemes expressing particular meanings (Goldberg (2003)). What is more, it is claimed that early child speech is an expression of the pre-linguistic categorisation of linguistic and extra-linguistic phenomena that proceeds along the same lines irrespective of the language (Langacker (1987), Bowerman (2003)). Consequently, it can be seen at first glance that despite different surface realisations, the same semantic notions are acquired both in English and in Polish. As Slobin (1970) has remarked: The rate and order of development of semantic notions are very constant across languages, regardless of the formal means of expression employed. Thus, this phenomenon has been studied by many researchers and was explained from a cognitive perspective long before cognitivism became influential in linguistics. For example, Brown (1973) and Bowerman (1973) have reported that the same semantic notions expressed the meanings of Stage I utterances across a wide range of languages. Ingram (1975a), Piaget (1963) and Sinclair-de Zwart (1971) have proposed that the semantic notions prevalent in Stage I speech reflect the cognitive distinctions among agents, their actions and objects acted upon that developmentalists regarded as the culmination of the sensori-motor period. When it comes to the classification of semantic categories adopted by various researchers, there are apparent differences between them. First of all, some categories are defined either in a broad sense, e.g. Schlesinger (1974) has applied a single category, modifier+x, to describe the relationship between the noun and the word, of whatever kind, modifying it; or they are divided into subtypes, e.g. the same Modifier+Head relationship can be expanded into Entity+Attribute or Possessor+Possessed (Brown (1973)) depending on the modifier preceding the noun. Yet another difference has to do with the fact that the same utterances can be classified either in terms of semantic notions, e.g. Locative action, e.g. Go to the table or Existence, e.g. This is a table (Bloom et al. (1975)) or semantic relations, i.e. respectively: Action+Locative and Demonstrative+Entity (Brown (1973)). There was a heated dispute in the 1970s whether to attribute knowledge of semantic notions only or of semantic relations to the child. According to 93
Bloom et al. (1975), [...] the cognitive representations of the relationship among objects and events are more global, not distinguishing between the respective semantic roles of the elements, but rather representing the entire relationship. Bowerman (1975) has been of a similar opinion and has suggested that the child has an internalised system of broad semantic notions rather than fine semantic relations. On the other hand, Schlesinger (1974, 1982) has proposed that all children’s utterances possess an internal semantic structure, i.e. that the child perceives the world in terms of agents, actions or objects and learns to map these categories directly onto the syntactic forms of his native language. Unfortunately, there is not enough empirical evidence to support either position. However, on the basis of my data composed of fragmentary word combinations, or even holophrastic utterances, I venture a statement that it is the semantic notions – not relations – that the child seeks to express, at least at the beginning of Stage I. In order to present the main semantic categories expressed in telegraphic speech in Polish and English I am going to use the classification system employed by Bloom et al. (1975). Since some of Bloom’s semantic notions are quite broad and, in my opinion, too general, I have decided to introduce several new semantic notions in order to provide a more detailed picture of the meanings that the child seeks to express. These new semantic notions are: Purpose, Order, Request for permission and Direction: Purpose
verb phrase referring to an idea of what the listener and/or the speaker are going to do Order verb phrase expressing an order or request Request for permission interrogative verb phrase, asking for permission Direction prepositional phrase describing the route of the subject or object movement Altogether, the proposed extended classification of semantic notions present in early child speech is as follows: Semantic notion Existence Attribution Negation Possession Locative action Locative state 94
Example Polish K (1;10) Osioł ‘donkey’ Z (1;8) Gorące ‘hot’ Z (1;8) Nie jajko ‘No egg’ K (2;2) Babci ‘Granma’ K (2;3) Malować tutaj ‘Paint here’ K (2;0) Tam zapałki ‘There matches’
English Jk (1;8) Car E (1;4) Crazy E (1;4) No cards E (1;4) Mamma’s J (3;1) Taking out there J (3;1) Under table
Purpose Order Action Request for permission Recurrence Direction
K (2;5) Robić bałwanki ‘Make snowmen’ K (2;1) Tutaj siedzieć ‘Sit here’ Z (1;8) Śpiewa ‘Is singing’ K (2;0) Rysować tu? ‘Draw here?’
E (1;8) Eat them
Z (1;9) Drugie jabłko ‘Another apple’ K (2;3) Do pokoju ‘To the room’
E (1;4) More chocolate
E (1;6) Jump E (1;6) Sleeping E (1;6) Taste it?
E (1;10) To market
As all the above examples come from my language corpus, it can be concluded that the same semantic notions are acquired at the beginning of linguistic development in English and in Polish. Although the order of their emergence varies from child to child, it has been observed that in all children studied, the first to be acquired were Existence, Attribution, Action and Negation. On the other hand, the notions outlined in this semantic classification are very general and the message expressed by them vague since they contain few or no grammatical morphemes that would be responsible for their further grammatical and semantic modification. In spite of this, they give us an approximate idea of the message intended by the child. They also support the cognitive theories of first language acquisition, according to which the onset of two-word utterances coincides with the end of the sensorimotor period during which the child organises the world in terms of things, actions and properties. These prerequisite cognitive skills underlie the semantic notions of agents, actions and attributes, and are invariable across languages (Slobin (1985)). Apart from the explanation of telegraphic speech offered by the cognitive theories of development, this phenomenon has also been explained from a completely different perspective, i.e. that of Universal Grammar, and in particular the Minimalist Programme, one of its most recent developments. In short, according to the Minimalist Programme, parameters, i.e. a system of rules which determine the way in which the language is structured, are contained within the lexicon, not in the structural part of grammar. Thus, all that children learning their mother tongue have to do is learn the lexicon of their language made up of both lexical categories, i.e. content words and functional categories, i.e. function words with their associated parameters. Radford (1990) has stated that children first acquire the lexical categories and their early speech is composed mainly of them with hardly any functional categories present,53 which means that they have not been acquired yet. 53
For a discussion of minimalism and first language acquisition see chapter 1.5.1.
95
However, the validity of this theory may be challenged on the grounds that it fails to explain the reasons for the very frequent omissions of content words at this stage. In fact, in my data content words tend to be omitted much more often than function words in the early stages of development, although with the passage of time this tendency is reversed. The reason for such a pattern of omissions may be that although the children at this stage have an internalised system of basic semantic notions, they have not yet acquired the basic syntactic rules which would enable them to form correct word combinations, i.e. sentences. Unfortunately, the Minimalist Programme is incapable of accounting for this phenomenon. All in all, traditional approaches to first language acquisition see the period of telegraphic speech as a result of making overgeneralisations about language structure, whereas cognitive theories state that at this stage of linguistic development children master the whole phrases, treating them as unanalysable wholes. Irrespective of the adopted approach, telegraphic utterances are quite ambiguous, giving rise to a wide variety of interpretations. 2.4.2. Ambiguity of sentences deprived of content words Before I proceed to analyse the sentences deprived of content words and the degree of their ambiguity, let us remark that the criteria applied both for selecting particular utterances and their interpretation are the same as those established for telegraphic speech. This means that I have chosen for my analysis only those utterances which were found systematically throughout the whole sample in all the children studied, even if their frequency was relatively low. Another thing is that I also intend to analyse only these utterances which have some scope for ambiguity. Moreover, I am going to stick to the principle of conventionality as well, which, here, again will limit the present discussion to consideration of semantic ambiguities that are not theoretically but practically possible. This means that I will restrict myself only to these expansions that are psycholinguistically plausible, i.e. ones that do not go beyond the child’s performance at that stage and, what is more, have been recorded either in the same child but in a different context or in other children in the same situation. Now I am going to introduce the semantic relations with a view to showing all the possible meanings that an incomplete utterance may express. I have decided to rely on employing semantic relations54 for this purpose since I am 54
Semantic relations can be perceived as the development and expansion of semantic notions and, in fact, the same utterances have been categorised either in terms of semantic categories or semantic notions. For example, Red car has been classified as Entity +Attribute by Brown (1973) and as Modifier +Head by Schlesinger (1971). In Bloom et al. (1975) and Leonard (1976), however, this utterance has been perceived as Attribution (De Villers and De Villers (1985)).
96
going to analyse the semantic structure of the utterances (which, I believe, can best be expressed in this way) rather than dwell on their general meaning (where semantic notions would suffice55). Another reason for choosing this option is that this analysis will go beyond Stage I. By the end of this stage, and even more so when it is over, the child is becoming more and more aware of the internal structure of the language and, because his competence finds its reflection in his performance, he attempts to form sentences consisting of more than two elements. The classification of semantic relations that I am going to present here and apply to my data is based on the system devised by Brown (1973). The order of these semantic relations has been established by myself on the basis of their frequency in the sample), whereas some of the terms for the patterns presented here together with the patterns themselves have been invented and employed by Brown. Innovative semantic notions that I have introduced have been marked with an asterisk. When it comes to the semantic relations, only some of them are present in Brown’s classification. These are: Demonstrative + Entity, Possessor + Possessed, Agent + Action and Action + Object.56 As regards the other semantic relations, I have established them on the basis of the recurrent patterns in the sample. As far as the novel semantic notions are concerned, they have been devised for the purpose of this analysis, as I believe that they will make it more detailed and present more effectively all the semantic subtleties of the incomplete children’s utterances. For example, Brown does not distinguish between Imperative, Action and Intention or Location and Direction, whereas they are definitely different on semantic grounds, as will be shown below. Here are my novel semantic notions and their definitions: Intention Direction Location
verb or verb + complement expressing willingness to perform an action adverb or prepositional phrase used to show the route of the subject or object movement. adverb or prepositional phrase that specifies the place of an action or the position of the subject
55
As far as the telegraphic speech is concerned, I relied on semantic relations because I was interested in the general semantics of the telegraphic utterances. 56 In his review of the present work Professor Szymanek has remarked that Object is disputable as a semantic category since it is – in its basic sense – a syntactic category because it defines a specific type of a grammatical relation. He has suggested replacing it with the notion of Theme/Patient, arguing that this would make the adopted terminological system more uniform. In principle, I agree with such a point of view, however, I have decided to preserve the concept of Object when discussing semantic relations as it has been used extensively by Brown in his seminal work (Brown (1973)), one of the milestones in the first language acquisition research.
97
State
verb denoting state, where no change of whatever kind with regard to the subject is involved Imperative verb in the imperative form expressing either a command or a request Recurrent noun noun whose recurrence is either commented on or asked for Non-existence negative phrase referring to the absence of a thing or person Rejection negative verb or phrase expressing lack of willingness to perform an action or agree to cooperate Non-action negative verb denoting the absence of an action Effect + Cause compound sentence, where the independent clause expresses effect, whereas the dependent one cause Donor Noun denoting the person who has given or offered something to the subject or object Now let me present a set of semantic relations made of semantic notions and expressed most frequently by children aged from 1;8 to 3;0, differentiated on the basis of my Polish and English sample: Semantic relation Demonstrative + Entity Intention* + Object Imperative + Dative + Object57 Possessor + Possessed Action + Object Agent + Action Action + Direction* Action + Location* State* + Location Imperative* (+Object) + Direction Intention +Recurrent noun* Vocative + Imperative Non-existence* + Entity Rejection* + Object More marginal relations: Agent + Non-action* Effect + Cause*
Example expansion ‘This is a boat’ ‘I want a cake/ I want to eat a cake’ ‘Give me the pencil’ ‘Mummy’s chair’ ‘Reading the book’ ‘She is reading’ ‘Went to the room’ ‘Jumping in the room’ ‘Is here’ ‘Go there/Put it there’ ‘I want another cake’ ‘Auntie, go’ ‘There is not a dog’ ‘I do not want the soup’
‘The egg did not break’ ‘I am washing the hands because they are white’
57 This is the most complete version of the pattern with the imperative, which can also function as an expansion of some utterances in its reduced version, namely without the dative, e.g. Paper for ‘Open the paper’.
98
Entity + Donor * Imperative + Instrument Action + Recurrent noun
‘Tiger from Ola’ ‘Draw with a pencil’ ‘I can see more hats’
The above expansions should be treated as utterances in which particular semantic relations are present in the deep structure of the language. For example, the utterance In the room in its surface structure only denotes Location but in its deep structure (Chomsky (1969)), 58 i.e. when provided with the intended meaning in the form of an expansion, it may stand for State + Location, i.e. ‘He is in the room’, or Action + Location, i.e. ‘He is jumping in the room’. Thus, this utterance expresses two kinds of semantic relations. Therefore, I can say that many of the utterances deprived of content words are structural homonyms, which means that they have the same structure but different meanings. At this point it is worth noting that some of these semantic relations are expansions of semantic notions used previously for the interpretation of telegraphic speech, e.g. Imperative + Direction: ‘Go there’ and Imperative + Dative + Object: ‘Give me the pencil’ are expanded forms of the semantic notion of Order which is claimed to express an order or request. Another example would be Action + Recurrent noun: ‘I can see more hats’ and Intention + Recurrent noun: ‘I want another cake’ are expansions of the semantic notion of Recurrence, e.g. More chocolate, which comments on or requests for the recurrence of an object or a person. These are obviously the most basic semantic relations sought to be expressed by children. However, they rarely occur in isolation in children’s speech in the way presented above but in various combinations with one another, which can easily be observed in the patterns presented in the next section. 2.4.2.1. The most common patterns of sentences devoid of content words from Polish and their expansions
In my analysis I will consider not only multi-word utterances but also single words which are very common as a medium of expression, especially in the early stages, denoting often the whole proposition, e.g. E (1;8) Water stands for the whole utterance ‘I want some water’. The issue concerning the status of single words has also been discussed in chapter 2.2. The procedure of forming expansions is the same as in the case of telegraphic speech, i.e. expansions a stand for what the child really meant by this particular utterance, whereas expansions b represent all the theoretically possible meanings 58
According to Chomsky (1969), each phrase or sentence is characterized by the surface structure, which is its form and deep structure, i.e. its meaning.
99
of a particular utterance, inferred by interpreting utterances of the same kind made by the same child, but in different circumstances, or by other children. The order of the utterances presented in this section has been established on the basis of their frequency in the sample, beginning with the most common patterns. Pattern I, made up of animate, inanimate and locative nouns, is prevalent in the one-word stage and very frequent in the two-word stage. Although the occurrence of this pattern is very common, it varies according to whether a particular child belongs to the group of referential or expressive children (Nelson (1975)). In my sample K. is a typical referential child, engaged in naming objects; therefore, in her speech names for objects predominate, constituting about fifty per cent of all the utterances On the other hand, Z. is an expressive child, more interested in social interaction and for this reason the percentage of nouns in her early speech is much lower, accounting for only about thirty per cent of all the patterns. Pattern II, subject + verb, and pattern III, verb + x, are quite common in the two- and three-word stage and their frequency is relatively invariable in all the three children, making up approximately thirty per cent of all the utterances. Patterns IV, prepositional phrase, V, noun + adverb and VI, infinitive + object, are common in the two- and three-word stage and each of them makes up from fifteen to twenty per cent of all the utterances. Pattern VII, another + noun, occurs regularly at the beginning of the two-word stage; however, when children enter the three-word stage, its frequency decreases sharply, constituting about ten per cent of all the utterances on the whole. Pattern VIII, negation + verb, can be found in approximately ten per cent of all the sentences as well. Patterns VIII, adjective + verb, and IX, negation + noun, are marginal in the corpus, constituting not more than five per cent of all the utterances. These are the most common patterns of sentences devoid of content words from Polish and their expansions: I. Nouns a) The examples belonging to this group come from children under the age of two; therefore the case distinctions are not preserved as the inflectional endings have not been acquired yet. 1. Inanimate nouns: Ołówek ‘Pencil’ Intended meaning a. ‘To jest ołówek’ ‘This is a pencil’ 100
Semantic relation Demonstrative + Entity
b. 1. ‘Rysuj ołówkiem’ ‘Draw with a pencil’ 2. ‘Chcę drugi ołówek’ ‘I want another pencil’ 3. ‘Daj mi ołówek’ ‘Give me the pencil’
Imperative + Instrument Intention + Recurrent noun Imperative + Dative + Object
2. Animate nouns: Ciocia ‘Aunt’ a. ‘To jest woda cioci’ Demonstrative + Possesor + Possessed ‘This is aunt’s water’ b. 1. ‘Ciociu, zjedz ciastko’ Vocative + Imperative + Object ‘Aunt, eat a cake’ 2. ‘Ciocia zażywa te tabletki’ Agent + Action + Object ‘Aunt takes these pills’ b) This group consists of examples from children who are older, i.e. over two, and in whose speech inflectional endings are present. 1. Animate nouns: Oli ‘Ola’s’ 2. a. ‘Poszłam do Oli’ ‘I went to Ola’s’ b. 1. ‘To jest tygrys od Oli’ ‘This is the tiger from Ola’ 2. ‘To jest Oli gazeta’ ‘This is Ola’s newspaper’
Action + Direction Demonstrative + Entity + Donor Demonstrative + Possessor + Possessed
At the time when children begin applying inflections, animate nouns used on their own occurred most frequently either in the genitive or the accusative. The genitive is so common because quite a large number of Polish prepositions require the noun in this case (Bąk (1989)). Another reason for this is that this case (genitive) denotes possession, which belongs to the earliest semantic notions acquired by children. At this point it is worth remarking that Brown (1973) differentiates between alienable and inalienable possessives. Alienable possessives describe a relationship between the possessor and the possessed that is not inseparable and can be changed, e.g. mummy’s cup, mummy’s chair, mummy’s towel, etc. Inalienable possessives specify an inseparable relationship between the possessor and the possessed because the latter is an inherent part of the former, e.g. mummy’s hair, mummy’s eyes, etc. According to Brown, alienable possessives are much more common than inalienable possessives and 101
the same has been observed in my data both from Polish and English: E (1;4) Horsie water for ‘Horsie’s water’, E (1;10) Mummy’s pen, Z (1;11) Rowerek Krzysia ‘Krzyś’s bike’, etc. 2. Inanimate nouns: Zupkę ‘Soup’ a. ‘Będę jadła zupkę’ ‘I will eat the soup’ b. ‘Daj mi zupkę’ ‘Give me the soup’
Agent + Action + Object Imperative + Dative + Object
3. Locative nouns: Pokoja ‘Room’ a. ‘Tatuś poszedł do pokoju’ Agent + Action + Direction ‘Daddy went to the room’ b. ‘Tatuś czyta w pokoju’ Agent + Action + Location ‘Daddy is reading in the room’ In 3, the adult form would be pokoj-u not pokoj-a, as the inflectional ending a is not correct because it marks the genitive case of neuter nouns. If the adult form is considered, it seems to have the same case ending -u intended for phrases expressing movement do pokoj-u ‘to the room’ (genitive) and then location w pokoj-u ‘in the room’ (locative). However, in Polish, the genitive ending is the same as the locative ending only in some instances and, as a result, the whole incomplete prepositional phrase is ambiguous. Such a homonymic case ending occurs when the noun stem ends in a palatalised consonant (Bąk (1989: 282)). Another example of this phenomenon would be genitive, do kuchn-i ‘to the kitchen’and locative, w kuchn-i ‘in the kitchen’. However, if the stem final consonant is hard, the endings for these cases are different: w bibliotec-e ‘in the library’ (locative), do bibliotek-i ‘to the library’ (genitive). Thus, this kind of ambiguity occurs only when the final consonant in the noun stem is palatalised. II. Subject + verb: Asia dała ‘Asia gave’ a. ‘Asia dała mi serek’ Agent + Action + Dative + Object ‘Asia gave me some cheese’ In this pattern there is no grammatical ambiguity because it is clear that the indirect and direct objects are missing but great semantic ambiguity can be observed because the number of nouns that could function as indirect objects here is practically endless. Likewise, four other patterns offer numerous interpretation possibilities: pattern III (verb+X) with a wide variety of subjects to choose from, pattern IV (prepositional phrase), which can be used with a broad spectrum of 102
agents and actions, pattern VI (infinitive+object) with the subject to be inferred and VIII (negation+verb) where the object is to be deduced. III. verb + X: Tutaj spała ‘Here slept’ a. ‘Kinga tutaj spała’ ‘Kinga slept here’
Agent + Action + X
X can be either Location, as here, or Object if the verb is transitive, e.g. Z (1;8) Nie dała całusa ‘Did not give a kiss’ for ‘Zuzia nie dała całusa’ ‘Zuzia did not give a kiss’, or Direction, e.g. Poszła do kuchni ‘Went to the kitchen’ for ‘Zuzia poszła do kuchni’ ‘Zuzia went to the kitchen’. In this pattern it is difficult to determine whether the omissions of the agent and object should be treated as the deletion of content words, i.e. nouns either common or proper, or function words, i.e. pronouns. In other words, it is not clear what children’s mental representations of the agent and object are like. However, on the basis of the data I have collected I may say that it is nouns that children resort to more often when forming their utterances, probably because of the fact that they mark the semantic and grammatical relations within the sentence more explicitly and overtly. Moreover, the number of possible agents is limited by the person, gender and number of the verb, which must be in agreement with the subject. Thus, it may be inferred that the child either means herself – ‘Kinga tutaj spała’ ‘Kinga slept here’ – or the 3rd person singular fem., e.g. ‘Mama tutaj spała’ ‘Mummy slept here’. This is not the case with English, where the same utterance: Here slept is much more ambiguous because the verb form is not marked for the person, gender or number. IV. Prepositional phrase: Do garnka ‘Into the pot’ a. ‘Wrzucam ziemniaki do garnka’ Action + Object + Direction ‘I am throwing the potatoes into the pot’ b. ‘Wrzuć to do garnka’ Imperative + Object + Direction ‘Throw it into the pot’ V. Noun + adverb (here or there): Tutaj misiu ‘Here teddy-bear’ a. ‘Postaw tutaj misia’ Imperative + Object + Direction ‘Put the teddy-bear here’ b. 1. ‘Narysuj tutaj misia’ Imperative + Object + Location ‘Draw a teddy-bear here’ 2. ‘Misiu tutaj spadł’ Agent + Action + Direction ‘The teddy-bear fell here’ 103
3. ‘Misiu tutaj siedzi’ ‘The teddy-bear is sitting here’
Agent + Action + Location
The same kind of expansions will also be in force for other adverbs, such as daleko ‘far’, blisko ‘near’ and wysoko ‘high’. However, I have decided to form the expansions with the adverb tutaj ‘here’ because it has the highest frequency in the sample, equalled only by the regularity of occurrence of the adverb tam ‘there’. When it comes to the subject of the above utterances, the noun misiu, it is the diminutive form of the noun miś ‘teddy-bear’, made by adding suffix -u, which is frequently used in the child language to create masculine nominative diminutive nouns. One of the children from my sample, K., relies on this suffix to a great extent: child diminutive dziadzi-u piesi-u koci-u królisi-u smoci-u dzidzi-u
adult diminutive dziadziuś, dziadzio piesek, piesiunio kotuś, kociuś Króliczek Smoczuś Dzidziuś
‘grandpa’ ‘doggie’ ‘kitten’ ‘bunny’ ‘little teat’ ‘dzidziuś’
The reason for employing the suffix -u to form diminutives is that it is the simplified and abbreviated form of the adult suffix -uś. According to Ingram (1992), final consonant deletion is a frequent phenomenon in the acquisition of the phonological system of language. Although Ingram conducted his study on a sample of English-speaking children, the same phenomenon may also be observed among children acquiring Polish, especially in the one-word stage. K (1;10): koczi for ‘kolczyk’, i.e. ’earing’, sima for ‘ślimak’ ‘snail’, kuli for ‘królik’, i.e. ‘rabbit’. Apart from the final consonant deletion, the reduction of consonantal clusters also takes place here, which is widespread in the process of the phonological acquisition of the language (Ingram (1992: 372)). In the above examples the stop, affricate or fricative and liquid /l/ or /r/ combinations have been reduced to stops, affricates or fricatives: lcz to cz in kolczi, śl to ś in sima and kr to k in kuli. VI. Infinitive + object:59 Trzymać kubek ‘Hold the mug’ a. ‘Chcę potrzymać kubek’ Intention + Object ‘I want to hold the mug’ 59 For other expansions of this pattern, see the section devoted to the analysis of telegraphic speech in Polish. Although as an example of a sentence deprived of content words this utterance is not ambiguous at all, I have decided to include it because it is very frequent.
104
VII. Another + Noun: Druga butelka ‘Another bottle’ a. ‘To jest druga butelka’ Demonstrative + Recurrent noun ‘This is another bottle’ b. 1. ‘Chcę jeszcze jedną butelkę’ Intention + Recurrent noun ‘I want another bottle’ 2. ‘Daj mi drugą butelkę’ Imperative + Dative + Recurrent noun ‘Give me another bottle’ VIII. Negation + Verb: Nie czyta ‘Is not reading’ a. ‘Mama nie czyta książki’ Agent + Non-action + Object ‘Mummy is not reading the book’ As stated in the theoretical part, the complexity of the utterance is directly proportional to the scale of omissions. Because the use of negation undoubtedly increases the complexity, here just the verb is present, whereas elsewhere I have noted the following utterances: Mama czyta ‘Mummy is reading’ or Mama zjadła ‘Mummy has eaten’, where at least the agent is present. IX. Adjective + verb: Białe myję ‘White wash’ a. ‘Myję ręce bo mam białe’ Effect + Cause ‘I am washing my hands because they are white’ b. ‘Myję białe talerze’ Action + Attribute + Noun ‘I am washing white plates’ This semantic relation occurred in two of the three children studied and is quite rare. Other examples would be: Mokre gotowałam ‘Wet cooking’for ‘Mam mokre ręce bo gotowałam’ ‘I have got wet hands because I was cooking’, and Nie gorące dotykam ‘Not hot touching’ for ‘Palniki nie są gorące, dlatego je dotykam’ ‘The burners are not hot that is why I am touching them.’ X. Negation + noun: Nie jajko ‘No egg’ a. ‘Nie ma jajka’ Non-existence + Entity ‘There is not an egg’60 b. 1. ‘Nie chcę tego jeść, chcę jajko’ Rejection + Intention + Object ‘I do not want to eat this, I want an egg’ 2. ‘Nie chcę jajka’ Rejection + Object ‘I do not want an egg’ 60
This is an example of telegraphic speech but I have decided to include it because it is the most frequent of all expansions.
105
3. ‘Jajko się nie rozbiło’ ‘The egg did not break’
Agent + Non-action
Having presented the most common patterns of sentences from which some content verbs have been deleted, let me now pass on to the English sample. 2.4.2.2. The most common patterns of sentences devoid of content words from English and their expansions
As in the case of the Polish patterns, the order of the utterances presented in this section has been established on the basis of their frequency in the sample, beginning with the most common patterns. Pattern I, made up of animate, inanimate and locative nouns, is the most frequent at the beginning of and throughout the two-word stage. Pattern II, noun + noun, pattern III, more + X, pattern IV, No + X and pattern V, noun + verb, are equally common during the two- and three-word stage, constituting about eighty percent of all the utterances. Pattern VI, verb + X, and pattern VII, prepositional phrase, are far less frequent, accounting for about ten percent of all the utterances. Lastly, pattern VIII, infinitive + noun, and pattern IX, preposition, are marginal in my sample. These are the most common patterns of sentences devoid of content words from English and their expansions: I. Nouns 1. Animate nouns: Dinosaur a. ‘I want to look at the dinosaur cards’ b. 1. ‘The dinosaur sits on this chair’ 2. ‘She is meeting the dinosaur’ 3. ‘This is the dinosaur’s towel’ 4. ‘Show me the dinosaur’
Intention + Attribute + Object Agent + Action + Location Agent + Action + Object Demonstrative + Possessor + Possessed Imperative + Dative + Object
2. Inanimate nouns: Paper a. ‘Open the paper’ b. 1. ‘She is reading the paper’
Imperative + Object Agent + Action + Object
3. Locative nouns: Table a. ‘She is sitting at the table’. b. 1. ‘She is going to the table’.
Agent + Action + Location Agent + Action + Location
II. Noun + noun: Horsie room a. ‘I can see the horsie in the room’ Agent +Action +Object + Location b. 1. ‘The horsie is going to the room’ Agent + Action + Direction
106
2. ‘The horsie is eating in the room’ Agent + Action + Location 3. ‘The horsie is painting the room’ Agent + Action + Object 4. ‘Horsie, go to the room’ Vocative + Imperative + Direction The ambiguity of this pattern and the preceding one is ensured by the lack of case endings. In Polish when children start to put two nouns together, they use them, on the whole, in the correct case, which excludes a variety of interpretations.61 For example, Dom mamusiu ‘Mummy house’ is limited to one interpretation: ‘Mamusiu narysuj dom’, i.e. ‘Mummy draw a house’ because the noun mamusiu ‘mummy’ has been used in the vocative case. Dzidziuś mamą ‘Baby mummy’ has been expanded into ‘Tam jedzie dzidziuś z mamą’ ‘There goes the baby with his mummy’ because mamą ‘mummy’ represents the instrumental case. Another example is Pingwin ogródka ‘Penguin garden’ for ‘Pingwin poszedł do ogródka’ ‘Penguin went to the garden, where ogródka ‘garden’ is a genitive case, which can be combined only with some prepositions. As far as the Polish data is concerned, I have found just one example in which inflectional endings have not been preserved in this pattern: Długopisik Kinga ‘Pen Kinga’ for ‘Kinga chce pisać długopisem’ ‘Kinga wants to write with the pen’. Because of this, this pattern is deprived of any ambiguity in Polish. III. More +X: More hats a. ‘I can see more hats’ b. ‘I want more hats’
Agent +Action +Recurrent noun Agent + Intention + Recurrent noun
In adult English the use of more is limited to the modification of countable and uncountable nouns. However, children often make category errors, i.e. they use more with verbs and even adjectives, so they are not aware of the fact that its use is so restricted. For example, E (1;8) said a few times: More dancing for ‘I want to dance more’. Thus, the construction of recurrence may comment on or request not only the recurrence of a thing, person, but also a process or activity. IV. No + X: No cards a. ‘There are not any cards’ b. ‘I do not want to play cards’
Non-existence + Entity62 Rejection + Object
61
For the discussion of noun cases see chapter III devoted to morphological iconicity. This is again an example of an expansion of a telegraphic speech utterance but because it exhibits the highest frequency in the sample, I have decided to include it here. 62
107
Apart from the fact that no and its Polish counterpart ‘Nie’ occur in word combinations, they can also be used on their own in holophrases and in such cases they stand for the whole sentence. For example, in English the negation no was found to have been used to express the following message: E (1;4) ‘I do not want to stay alone’. The corresponding example from Polish would be nie ‘no’ for K (2;0) ‘Nie płucz śliniaka’ ‘Do not rinse the bib’. V. Noun + verb63: They pick up a. ‘They are picking the crayons up’
Agent + Action + Object
VI. Verb + X: Reading book a. ‘Daddy is reading a book’
Agent + Action + X
As in the Polish sample, X can be Object, Direction or Location, but because Object is the most frequent I have used it in the expansion. It may seem that in English there should not be a clear-cut distinction between the agent + action and the action + object pattern because nouns lack inflectional endings and the word order, although in the majority of cases correct, sometimes is not preserved. However, in my data I have not found a single example of incorrect word order in sentences of this kind and, consequently, there is no confusion as to whether a given noun is the agent or the object. Thus, it may be concluded that the rules concerning the correct word order are more strictly observed than might seem. Therefore, the statement may be put forward that the billiard ball model, proposing that the object follows the verb, in the same way as the reception of energy follows its flow, is deeply rooted in the children’s cognitive system. VII. Prepositional phrases: Under table a. ‘Put it under the table’ Imperative + Object + Direction b. 1. ‘It is under the table’64 State + Location 2. ‘I have put it under the table’ Agent + Action + Direction 3. ‘I am sitting under the table’ Agent + Action + Location Under is one of the prepositions ensuring a substantial degree of ambiguity, as it can be used both to specify the direction of movement and the 63
This construction is very rare in English. Moreover, the kind of semantic ambiguity for this pattern and also the next one, i.e. VI, is of the same kind as for the Polish utterances consisting of subject + verb or verb + X. 64 This is, in fact, an expansion of telegraphic speech but I have decided to include it here so as not to disrupt the entire process of the disambiguation of this particular utterance.
108
location, as is the case with the majority of prepositions which can refer both to movement and position, e.g. above, over, below, beneath, between, off etc. However, when it comes to the prepositions in and on, which are most often used by children, they can only denote position because in order to express movement their counterparts, into and onto have to be used, so they are far less ambiguous when used in prepositional phrases. VIII. Infinitive + noun: Bite it.65 a. ‘I want to bite it’. b. ‘Dolly is biting it’
Intention + Object Agent + Action + Object
IX. Prepositions: On a. ‘I have put the shoe on’ b. ‘The shoe is on’
Agent + Action + Object + Direction Agent + State + Location
Other holophrases consisting only of prepositions from my corpus would be, e.g. Up for ‘I am moving the blocks up’ and Out for ‘I will get it out’. In Polish such examples are non-existent because the English verb + particle or preposition construction has no immediate equivalent. The meaning expressed by particles or prepositions in English is denoted in Polish by prefixes, e.g. wyciągnąć ‘take out’ or prepositional phrases like do góry ‘up’ or simply by verbs, Ubrałam buta, which are counterparts of the verb + particle phrase ‘I have put the shoe on’. 2.4.2.3. Analysis of sentences devoid of content words on the basis of Polish and English data
Now, I will verify the above data in the light of the theoretical assumptions concerning the origin of the omission of contentives, included in the theoretical part of this chapter (ch. 2.3.2). First of all, as the data examined has shown, the majority of the utterances seem to be the result of overgeneralisations that children make about the structure of language. In fact, most of these word strings resemble nonsentence fragments allowed in adult discourse as either answers to specific questions or explanations and clarifications of what has just been said. In order to show the direct analogy between the adult model and the children’s imitations as well as the inappropriate application of this model, let us imagine the following exchanges that could have taken place in adult language:
65
This utterance, as in the case of the telegraphic speech utterances, can be perceived as a totally complete and perfectly correct imperative sentence.
109
Question Who is she meeting? Where is she going to? What can you see? Where shall I put it? What is she doing?
Non-sentence answer66 The dinosaur. Table. More hats. Under the table. Reading the book.
As the data under discussion shows, children repeat non-sentence fragments, heard in adult language, treating them as fixed phrases, and resort to them irrespective of linguistic context. Thus, it can be confirmed that there exists a very strong link between elliptical adult language and children’s early utterances in the sense that children perceive the existence of omissions in the adult language but at the same time use them indiscriminately, treating them as unanalysable chunks. Secondly, as far as the principle of complexity is concerned, the data confirms the direct connection between sentence complexity and the degree of difficulty experienced by the child in producing or reproducing a particular utterance. This phenomenon results from performance errors, i.e. the inability to include all the necessary sentence elements. When referring to the factors mentioned in chapter 2.3.2.2, which contribute to the increase in the number of deletions in the data, it is the use of negation and the noun modifier that pose the major problems. It means that children tend to omit content words if a particular utterance contains a noun modifier or negation. However, if the same or a similar utterance does not include a noun modifier or negation, it is produced correctly and all its constituents are present (the modifiers have been written in the bold type): Utterances containing modifiers contentives omitted intended meaning 1. E (2;0) I wanna sharp the black crayon ‘I want to sharpen the black crayon.’ 2. E (2;6) I don’t give you a blue Cup ‘I won’t give you a blue cup’ 3. E (1;6) No cards I don’t want/play ‘I do not want to play the cards’ 4. K (2;2) Nie widać Tatusia ‘Nie widać tatusia’ ‘You cannot see’ ‘daddy’ ‘You cannot see daddy’ 5. K (2;2) Ten kocyk Daj mi ‘Daj mi ten kocyk’ ‘This blanket’ ‘Give me’ ‘Give me this blanket’ 6. K (2;3) Czyste mam Ręce ‘Mam czyste ręce’ ‘I have got clean’ ‘hands’ ‘I have got clean hands’ 66
110
These non-sentence fragments come from my sample.
Corresponding utterances Intended meaning without modifiers with no elements omitted 1. E (1;8) I wanna take boots off 3. E (2;0) Wanna do stamps 2. E (1;4) I get the cards 4. K (2;1) Zaświecić światło ‘To turn the light on’ 5. Z (1;8) Łyżeczkę dam ‘I’ll give the teaspoon’ 6. K (2;2) Misiu ma budzik ‘The teddy-bear has got an alarm clock’
‘I wanna take my boots off’ ‘I wanna do stamps’ ‘I’ll get the cards’ ‘Trzeba zaświecić światło’ ‘You should turn the light on’ ‘Daj mi łyżeczkę’ ‘Give me the teaspoon’ ‘Miś has got an alarm-clock’ ‘The teddy-bear has got an alarmclock’
It is interesting to remark that in utterance 5 the child has used the first person singular future form of the verb – dam ‘I’ll give’ – instead of the imperative, daj ‘give’. The reason for such a use of this verb, which can be found in other utterances as well, is that children very often engage in the process of imitation (Jespersen (1971), Scollon (1976)). As the form dam ‘I’ll give’ has been used frequently by the mother of the girl when giving her the things she asked for, the girl just imitated this form and relied on it also in making requests. When trying to explain this case in the light of the latest cognitive theories of language acquisition, it has been argued that young children are less creative than was previously thought (Lieven et al. (1997), Tomasello (2000), Dąbrowska (2002)). Lieven et al. (1997) conducted research into the acquisition of multiword utterances by a child at the age of 2;1 and, as a result, found that as many as 63 per cent of all utterances are repetitions of utterances already heard before by the child. For this reason Croft (2001) has concluded that it is grammatical constructions, not categories or relations that are the basic units of language. This finding has been confirmed by Tomasello (2000) and Dąbrowska (2000b), who have argued that children’s earliest utterances are made up of single words or indivisible and invariable phrases, such as there is not, and what’s this? Dąbrowska and Tomasello have explained this by claiming that children hear contextualized utterances and acquire not the separate words constituting these utterances but the entire phrases or utterances together with their meaning. Thus, when the girl from my sample hears her mother saying Łyżeczkę dam ‘I will give the teaspoon’, she learns the whole phrase, treating it as an inseparable unit to be used in situations when people give things to each other, irrespective of who is the initiator of the action. 111
When it comes to the individual differences and pragmatic factors (chapter 2.3.2.3), several variations have been noticed, both from child to child and from one situation to another. For instance, one of the children from the Polish sample, K., concerned with providing names for the objects from her immediate surroundings, used nouns most extensively in the initial stages of her linguistic development, while other sentence elements, such as verbs or adjectives, were omitted. In contrast, another one, Z., relied both on nouns and verbs, commenting on what was going on. Moreover, it has been noticed that the context of the conversation has an immediate impact on the use of particular semantic relations, e.g. when playing with parents lots of imperatives were used by means of which children expressed their requests. On the whole, in the traditional view, the reasons for the omission of content words are both of linguistic and extra-linguistic in nature. First of all, children make overgeneralisations about the language input. Secondly, the children’s cognitive and linguistic system is incapable of processing and producing more complex utterances. Lastly, both individual differences and context do influence the pattern of omissions.
2.5. Semantic analysis of sentences devoid of content and function words in Polish and English In this section I will analyse the children’s utterances which are devoid of functors and contentives from the semantic point of view, i.e. I intend to examine the range of all possible meanings that a particular utterance may express. Therefore, I need to introduce several novel semantic notions that would communicate precisely the intended meaning of the children’s incomplete utterances. For this purpose, I am going to divide the broad concept of Action, which has been used by the majority of researchers (Brown (1973)) to denote any action, into finer semantic categories: Future action Continuous action Regular action Necessity Proposal Plan
action referring to future action in progress at the moment of speaking action taking place regularly action that must or should be done. action that has been suggested. action that has been planned.
Another semantic notion that I am going to introduce here and use in the analysis is that of Beneficiary. It was created by Brown (1973) to talk about the 112
person who will benefit from the agent’s action, e.g. in the sentence I have bought it for you, you is the Beneficiary. At the beginning of this analysis, one thing that both languages have in common should be noted. On the whole, the first pattern of incomplete utterances to emerge consists of nouns only, where the noun, depending on its category and form, may denote both in Polish and in English any of the following: Instrument (Pencil for ‘Draw with a pencil’), Recurrent noun (Cake for ‘I want another cake’), Object (Cake for ‘I want a cake’), Possessor (Mummy for ‘Mummy’s chair’), Vocative (Auntie for ‘Auntie, go’), Agent (Auntie for ‘Auntie takes these pills’), Donor (Oli for ‘This is the tiger from Ola’) and even Location (Table for ‘At the table’) and Direction (Table for ‘To the table’). Thus, the range of semantic interpretations of such utterances is quite wide and if compared with the examples of telegraphic speech expansions consisting merely of single nouns, it can be concluded that the former are much more ambiguous than the latter. If the above semantic classification is applied to the nouns standing for the telegraphic speech utterances, it can be seen that they stand for Entity (Car for ‘This is a car’) or Possessed (Car for ‘This is my car’), and when they are used as parts of incomplete prepositional phrases they may represent also Location (Car for ‘It is in the car’) or Beneficiary (Daddy for ‘For Daddy’), so the range of meanings that they express is much narrower. Because, as has been shown in the course of this analysis, there are some patterns which, out of context, can be classified either as sentences deprived of contentives or functors, let us see whether the same phenomenon may be observed for them. These are: Verb + Object/Location/Direction, Noun + Verb, Negation + Noun and Noun + Noun, Here/there67 + Noun and Adverbial/prepositional phrases. When it comes to the Infinitive + Object combination (’Eat it’/Jeść to), in Polish telegraphic speech expansions it can have the following interpretations: Intention + Object (‘Muszę to zjeść’ ‘I must eat it’), Agent + Request68 + Object (‘Czy mogę to jeść?’ ‘Can I eat it?’), Agent + Future action + Object (‘Będę to jeść’ ‘I will eat it’), Necessity + Object (‘Trzeba to jeść’ ‘This should be eaten’) and Imperative + Object (‘Masz to jeść’ ‘You are to eat it’). All these semantic relations are also present in the English sample. Apart from them, there are also: Agent + Continuous action + Object, also in the 67
In this pattern here and there function as prolocatives. I have already used this term when dealing with the semantic interpretation of telegraphic speech in the preceding chapter. Then I discussed the meaning of incomplete utterances such as Draw here? and I decided that the semantic notion they express is Request. Expansions of such sentences: as Can I draw here? denote the same meaning, i.e. that of a request. Moreover, the agent is added; hence the pattern: Agent + Request. 68
113
interrogative form (‘I am eating it’ or ‘Are you eating it?’), Agent + Regular action + Object (‘He eats it’), Proposal + Object, (‘Let us eat it’), and Agent + Plan + Object, (‘I am going to eat it’). Strangely enough, the pattern Infinitive + Object, when deprived of content words, i.e. in this case the subject (Read a book), is most often expanded into only two patterns in English – the ambiguous Agent + Action (+Object) (‘She is reading a book’) or Agent + Intention + Object (‘I want to read it’) – and just one in Polish: Agent + Intention + Object (‘Chcę to czytać’ ‘I want to read it’). The Noun + Verb pattern of telegraphic speech has different realisations in Polish and in English. In Polish (Zuzia piła ‘Zuzia was drinking’) it is always Agent + completed Action (Zuzia wypiła ‘Zuzia drank’) or Agent + future Action (Zuzia będzie piła ‘Zuzia will drink’), whereas in English this construction (Daddy drink), although rare, has many interpretations. It can be Agent +Action (‘Daddy is drinking’), Vocative +Agent +Action (‘Daddy, I am drinking’), Vocative + Question (‘Daddy, are you drinking?’) and Vocative + Imperative (‘Daddy, drink’). If this combination is considered in terms of the content words omitted, both in Polish and in English, it will always be the object, expressed by the common or proper noun, that is missing and therefore the semantic interpretation of this sentence in the two languages will always be: Agent + Action + Object (‘Zuzia piła mleko’ ‘Zuzia was drinking milk’, and ‘Daddy is drinking the coffee’). Negation + Noun combination (No cards/Nie jajko ‘No egg’) in telegraphic speech in both languages is used to denote Non-existence + Entity (‘There are not any cards’/’Nie ma jajka’ ‘There is not an egg’), whereas in the sentences deprived of content words the meaning that is most often expressed is Agent + Rejection + Object (‘I do not want an egg’). Apart from that, in Polish it is also used to denote Rejection + Intention + Object (‘Nie chcę tego, chcę jajko’ ‘I do not want this, I want an egg’) and Agent + Non-action (‘Jajko się nie rozbiło’ ‘The egg did not break’) The Noun + Noun pattern is definitely English-specific as in Polish it does not occur as an example of telegraphic speech. In the utterances in which content words are deleted this pattern is devoid of ambiguity because the nouns which appear in it are used in the specific case, which eliminates a multiplicity of interpretations. For example, in the utterance K (2;3) Dzidziuś nog-ą ‘Baby leg’, the noun nog-ą represents the instrumental case; therefore, the choice of verbs is limited to those which take the noun with the instrumental case ending, such as rusza ‘is moving’, fika ‘is kicking’, etc. Because of this, only the English pattern will be studied, which has turned out to be equally ambiguous in both cases. In telegraphic speech (Bear hat) it can denote the following: Demonstrative + Possessor + Possessed (‘That is a bear’s hat’), Demonstrative + Attribute + Noun 114
(‘That is a bear hat’), Demonstrative + Entity + Entity (‘That is a bear and a hat’), Demonstrative + Entity + Beneficiary (‘That is a hat for the bear’), Subject + State + Object (‘The bear has got a hat’) and Demonstrative + Entity + Location69 (‘That is a bear in the hat’). With regard to sentences devoid of content words there are five expansions of this string, presenting five different semantic interpretations.70 The present corpus shows that Here/there + Noun pattern (Tutaj misiu ‘Here teddy-bear’) is also specific to one language only, this time Polish. Here and there function in the above examples from the Polish sample as ‘pro-locatives’, i.e. they replace complete prepositional phrases with the locative function. This indicates that their absence from the English sample is caused by the fact that English prepositional phrases are easier to acquire than Polish ones. The reason for this may be again the great simplicity of the English inflectional system as compared with Polish, where the child not only has to learn the right preposition but also the correct inflection of the noun. Although the Here/there pattern can also be found in the English sample, its frequency is too low to show the range of all its possible meanings on the basis of this data. With regard to Polish telegraphic speech, this pattern has only one meaning, Agent + State + Location (‘Tutaj jest misiu’ ‘Here is teddy-bear’), whereas in sentences deprived of content words it has four different expansions: Imperative + Object + Direction (‘Połóż misia tutaj’ ‘Put the teddy-bear here’), Imperative + Object + Location (‘Narysuj tutaj misia’ ‘Draw a teddy-bear here’), Agent + Action + Direction (‘Położę misia tutaj’ ‘I’ll put the teddy-bear here’) and Agent + Action + Location (‘Narysuję misia tutaj’ ‘I will draw a teddy-bear here’). Lastly, as regards prepositional phrases (Pod stołem/‘Under the table’), they denote, both in Polish and English, only Entity + State + Location (‘To jest pod stołem’/‘It is under the table’) as expansions of telegraphic speech, so they are not ambiguous at all. The unequivocal interpretation of this pattern results from the fact that of all the function words, the verb to be/być is the only verb that can be used in the expansion of the above utterance. As utterances deprived of content words, prepositional phrases are not ambiguous in Polish either, since they denote only one semantic relation: Agent + Action + Location: (‘Pies leży pod stołem’ ‘The dog is lying under the table’). The reason for this is that in Polish the majority of prepositions require different noun cases depending on whether the prepositional phrase denotes Movement or Location. Thus, the noun occurs in the accusative case when the prepositional phrase refers to the movement of a person or thing, e.g. Chowamy się (przed deszczem) pod dach ‘We are hiding under the roof’, where dach ‘roof’ is in the accusative case. 69 70
Here by in the hat I mean ‘placed inside the hat’, not ‘wearing the hat’. See pattern II above.
115
However, when the prepositional phrase describes the location of a person or thing and it answers the question where?, the prepositions are followed by a noun in the locative or instrumental case, e.g. Jesteśmy pod dachem ‘We are under the roof’, where dach-em is the instrumental case (Bąk (1989: 184)). In English, apart from Agent + Action + Location, the prepositional phrase can be expanded into two more patterns because of the lack of the noun inflections: Imperative + Object + Direction (‘Put it under the table’) and Agent + Action + Direction (‘We are putting it under the table’). Now let us have a look at the table below, where the degree of ambiguity of telegraphic speech and the utterances deprived of content words has been presented.
Pattern Noun Infinitive + Object Noun + Verb Negation + Noun Noun + Noun Here/there +Noun Prep. Phrases
Possible expansions of Possible expansions of sentences telegraphic speech without content words Polish English Polish English 4 4 9 9 5 9 1 2 2 4 1 1 1 1 3 1 – 6 – 5 1 – 4 – 1 1 1 3
As shown above, in Polish the ambiguity of utterances deprived of content words is greater than the ambiguity of those being examples of telegraphic speech, while in English the opposite is the case. As regards Polish, the patterns: Noun, Negation + Noun and Here/there + Noun convey more possible meanings if content words have been omitted. Infinitive + Object and Noun + Verb patterns are more ambiguous for telegraphic speech, whereas the remaining pattern, i.e. Prepositional phrase, is ambiguous neither when content nor when function words are omitted. In English the Noun pattern, Noun + Noun as well as Prepositional phrase are more ambiguous when deprived of content words, whereas Infinitive + Object and Noun + Verb convey more meanings as examples of telegraphic speech. Unlike in Polish, the Negation + Noun sequence is not ambiguous for the either pattern. According to what has been predicted at the beginning of this chapter, in Polish the omission of content words, capable of expressing a wider variety of concepts, entails more ambiguity than deleting function words limited in their number and meaning. However, in English the situation is different because of 116
the fact that in English basic word forms do not end in grammatical morphemes but possess a zero ending. Consequently, it is possible to talk about adding various inflectional morphemes to them and in this way changing their meaning. This is why it is possible to assign so many different meanings to telegraphic speech utterances in English. In Polish, because inflectional morphemes are not added but replaced with the required ones when forming a new grammatical category, the number of telegraphic speech expansions is limited, as is their ambiguity.
Conclusion To sum up, all the omissions, of both content and function words, that have been discussed in this and the preceding subchapters are not purposeful because they are by-products of the developmental restrictions and processing limitations of the children’s cognitive system. If the emergence of patterns devoid both of function and content words is traced chronologically, it can be easily remarked that they are an expression of the radial strategy of language acquisition. Thus, children first master prototypical, i.e. radial speech categories, e.g. concrete nouns naming objects from their environment, or light verbs. They then acquire basic sentence patterns only to expand them into more sophisticated structures until they eventually reach adult competence. For example, the majority of the children studied first acquire semantically simple nouns, which they then begin to combine with light verbs, prepositions, and other nouns. Finally, in a stage that is beyond my considerations, children combine all the sentence elements, hitherto used in a limited number of configurations, to form grammatically correct utterances. Moreover, as was stated in the preceding sections there are two reasons for the phenomenon of omissions in children’s speech. First of all, according to cognitive linguistics, in the initial stages of linguistic development children are less creative than was previously thought. Therefore, they master entire phrases without subjecting them to any grammatical analysis and use them indiscriminately, irrespective of the context. Secondly, at the onset of linguistic development children overgeneralise, i.e. simplify, to a large extent the rules of phrase and sentence formation, which leads to the production of elliptical utterances. Children resort to the simplification strategy quite often, which is responsible for the creation of the majority of forms non-existent in adult language.71 71
The strategies of simplification that children rely on are discussed in detail in chapter III, which is devoted to the issue of morphological and syntactic iconicity.
117
These two strategies, i.e. acquiring entire language chunks and simplifying grammatical rules, result in making children’s utterances ambiguous outside of their context, although, if the pragmatic context is ensured, they are well understandable for close family members of the child. Another consideration is that incomplete English utterances are more ambiguous than Polish ones due to the impoverished system of English inflectional morphology.
118
CHAPTER III
Morphological and syntactic iconicity in child language Introduction In this chapter I will show how iconicity manifests itself in child language. As the phenomenon of iconicity goes back to the classification of linguistic signs, my first aim will be to give various definitions of the linguistic sign in the system of both verbal and non-verbal communication from an extralinguistic and linguistic perspective to provide a broad background for the present discussion. I will present a distinction between motivated and arbitrary signs, defining three basic types of signs encountered in any system of communication, namely index, icon and symbol. Secondly, I will discuss the icon as linguistic sign and the issue of both morphological and syntactic iconicity in children’s speech. On the whole, the aim of this chapter is to prove that child language is much more iconic than adult language as far as both its morphology and syntax are concerned. With regard to the morphological analysis to be carried out, I am going to adopt the framework of Natural Morphology, which will serve as a tool for analysing both inflectional and derivational processes discernible in children’s speech. I will prove that the internal structure of words coined by children is motivated, i.e. iconic. When it comes to the syntactic analysis, my aim is to show that the internal structure of many sentences produced by children exhibits a close similarity with the meaning expressed by them, i.e. it is iconic as well. Now, I will discuss the development of the notion of sign in twentieth century linguistics, beginning with the definition of the linguistic sign put forward by Ferdinand de Saussure and ending with the semiotic theory of the linguistic sign advanced by Charles Peirce. To ensure a more comprehensive look at the phenomenon of the linguistic sign I am also going to include some extralinguistic considerations concerning the system of human communication.
119
3.1. Historical background Iconicity in language is a notion that has recently gained immense popularity (Dirven (1999), Kardela (1994), Dąbrowska and Kubiński (2003), Tabakowska (2001)). Generally speaking, iconicity allows us to perceive a similarity between the linguistic form and the object that this form denotes. For example, onomatopoeic expressions are iconic, as they imitate a wide variety of sounds. Thus, onomatopoeic words, such as twitter and knock-knock are linguistic icons because they show a relation of analogy with the actions they stand for. The appeal of linguistic iconicity has been spurred by its widespread application for research in the field of cognitive linguistics, which regards language as a system of communication consisting of three types of linguistic signs: indexes, icons and symbols (Peirce (1965)). Each of these signs differs from the others with respect to the relation between its overt structure and the meaning that it expresses. However, before I proceed to discuss these modern cognitivist tenets of language structure, I will relate to the traditional theory of the linguistic sign put forward by the father of structuralism, Ferdinand de Saussure, at the beginning of the twentieth century. Ferdinand de Saussure (de Saussure (2002)) claimed that the linguistic sign is made of the union of two parts: the concept and the sound image. The concept is understood as the mental image created in the mind of the language user, whereas the sound image is defined not as a physical sound but as a psychological imprint, the impression that it makes. The union between the concept and the sound image is a close one, as one part will instantly conjure up the other. Saussure argues that, for example, there are various words to express the concept ‘tree’ in various languages. Being a speaker of a certain language, the sound image for a tree in that language will automatically conjure up the concept ‘tree’. Thus, the meaning of any linguistic sign is found in the association created between the sound image and the concept. The sound image has been defined by Saussure as signifier and the concept as signified. Saussure states that the main characteristic feature of the linguistic sign is its arbitrariness. This means that the relationship between the signifier and signified is not motivated in any way, but is based exclusively on convention. Thus, there is nothing in the signifier that would determine its use to express a particular signified. To illustrate this phenomenon, Saussure says that there are different words in different languages for the same thing, for example dog is ‘dog’ in English, ‘perro’ in Spanish, ‘chien’ in French, ‘Hund’ in German. Saussure also insists that there is not a single non-arbitrary sign in language. This claim is difficult to uphold in the light of the fact that meanings of many words are definitely motivated, i.e. there is a logical relation between the signifier and the signified, as for instance in onomatopoeia words where the 120
signifier is an expression of the sound that the signified makes. Other examples of this kind would be, for instance, expressions connected with making phone calls: we say that we hang up the receiver and dial the number. The verb hang up is semantically motivated because it describes the action that actually takes place, i.e. when we hang up the receiver, we place it in a hanging position. Similarly, when we dial the number, we use for this purpose the telephone dial, so in this case the choice of the verb is not arbitrary either. Also, some words, that in the past demonstrated a meaningful relation between meaning and form, often assume an arbitrary character with the passage of time. Nowadays, we put down the receiver much more often than hang it up and we usually press buttons to make a phone call as telephones are rarely equipped with dials. Consequently, the verbs hang up and dial have assumed an arbitrary meaning, but from a diachronic perspective they are semantically motivated. However, Saussure, like all structuralists, focuses on a synchronic analysis of language, completely ignoring the diachronic perspective in the analysis of linguistic phenomena. He claims that all linguistic items, even those exhibiting an obvious resemblance with the objects or actions that they denote, are arbitrary because they are based on community agreement. In other words, he is of the opinion that communities agree on fixing or changing the relationships between signifiers and signifieds; however, he does not account for this fact in any way. Thus, he offers no explanation for the fact that the meaning of some words is semantically motivated. This evident gap in the theory of the language sign, a failure to explain the nature of some meaningful relations between the signifier and the signified, was eradicated with the advent of cognitive linguistics. As has been remarked above, cognitive linguistics dismisses the idea that all linguistic signs are arbitrary. In the next subchapter the nature of the linguistic sign will be discussed from a cognitivist perspective which proposes the existence of motivated signs in language beside arbitrary ones. 3.1.1. Arbitrary and motivated signs As has already been stated, the notion of the sign has been expanded and developed as a result of the growing popularity of cognitivism, which relies on the study of signs to a great extent. The study of signs is called semiotics, a name created from the Greek word semeioticos, meaning ‘concerning the sign’ (Dirven and Radden (2001)). Semiotics deals with the analysis of both verbal and non-verbal forms of communication between people and animals, such as, for example, human language, gestures, the significance of clothes that people
121
wear, the distance between speakers, or situations in which animals bare their teeth, roar, or perform dance rituals in order to pass a certain message. Evidently, the human system of communication represents the most sophisticated and highly developed system of signs. Humans are social beings, feeling an overwhelming need to communicate with one another, using a wide range of means for this purpose. We express astonishment by raising our eyebrows, we can use our finger to ‘draw’ the shape of a particular object in the air, or express our thoughts by means of words. All these media of expression are signs of something for us. Generally speaking, the sign may be interpreted as a particular form that represents ‘something else’ that should be understood as its meaning (Tabakowska (2001)). All the above media of expression represent the three possible kinds of signs: indexes, icons and symbols. Before I proceed to make a classification of these signs from a linguistic point of view and analyse in what way they structure the language, I am going to describe them from a more general, i.e. an extra-linguistic, perspective (RudzkaOstyn (1988)). To begin with, an index is a sign that directly points to the object existing in its immediate proximity, without describing it in any way. Tracing the etymology of this term, it turns out that in Latin index means ‘index finger’, which traditionally is used to point to things, people or places in the same way as the indexical sign is. The most obvious example of an index would be a signpost placed at the side of the road, showing the direction to and the name of the nearest town. For instance, the signpost Dover should be interpreted as: ‘You should go this way in order to get to Dover’. Another example of an indexical sign would be the facial expression of raising one’s eyebrows, mentioned above, pointing to the emotional state of the person, in this case, astonishment. An iconic sign is regarded as the image of an object of visual, phonetic or other nature, an image that can be subject to perception. The term icon derives from Greek, in which eikon means ‘image’. The iconic sign resembles the object that it represents. For instance, a road sign in the vicinity of a school warning drivers not to drive too fast, depicting two children crossing the street is an iconic sign. However, this icon provides nothing more than a very rough sketch of reality, as at any given moment there might be only one child crossing the street, or a great many children. Despite this, the meaning of this iconic sign is well pronounced, as it stresses the importance of driving slowly for the sake of the children’s safety. The above mentioned gesture of ‘drawing’ shapes in the air is also an example of an icon because it points to the resemblance between the object ‘drawn’ in the air and the real one. Although indexes and icons are two different kinds of signs, they share one common feature: the relationship between their form and the represented object 122
is a natural one. This means that it is motivated, i.e. it can be logically accounted for. In the case of an index the relation between the meaning and the form is based on either proximity or causality. It is possible to talk about proximity in the case of, for example, gestures or road signs because in these cases a very close relationship between the signifier and the signified can be observed, i.e. a kind of semantic contiguity. As far as causality is concerned, it can be said that the smoke, which indicates fire, is an index, as it is referred to its referent by causality, i.e. it is caused by it. With regard to an icon, the relationship between the signifier and the signified is more complex as the observer must become aware of the iconic relationship of similarity. The icon may be very similar to the object, as in pictures depicting saints, or it may be more abstract, as in the case of schematic silhouettes of cars or planes depicted in road signs. As opposed to the index and the icon, the symbol is a sign whose form and meaning are related only by convention, i.e. arbitrarily. An example of the symbol would be a road sign presenting an upside-down triangle, whose meaning is ‘give way’. There is no natural relation between the form of the sign and its meaning. This relation is purely conventional. The same can be said about symbols for currencies, national flags and, unquestionably, about the majority of linguistic signs. 3.1.2. Motivated and arbitrary signs in language Undoubtedly, the nature of language is largely symbolic, because, in the majority of cases, the relation between the form and the meaning is based neither on proximity nor on contiguity but on convention. This was discovered and described by Ferdinand de Saussure. On the other hand, in this intricate system of symbols referred to as language, the existence of motivated linguistic signs, such as indexes and icons can also be observed. This has been recognised by cognitive linguistics, which has made use of the linguistic definition of an index and an icon (Taylor (2002)). The semiotic theory of Charles S. Peirce (1965) has been adopted by cognitive linguistics as the superordinate framework for the discussion of the nature of the linguistic sign (Galeas (1998)). According to Peirce, a sign consists of a signans, which expresses a signatum. The signatum, in turn, stands for what is expressed in the sign. As far as the very basis of this theory is concerned, a direct analogy can be drawn between it and the definition of the linguistic sign put forward by Ferdinand de Saussure. Thus, signans can be equated with signifier, i.e. the form, whereas signatum stands for signified, i.e. the meaning. However, this is the only point of convergence between these two theories. As has been stated above, de Saussure advocates the arbitrary character of the
123
linguistic sign, whereas Peirce focuses on the opposition between natural and conventional. In what follows I will present the semiotic theory of Charles S. Peirce in detail and show how it has been employed to explain the nature of some motivated linguistic phenomena and provide a conceptual framework for Natural Morphology. Peirce defines a symbol as a sign that refers to its object through a conventional or habitual link. This link that connects a symbol to its referent must be known both by the speaker and listener in order to be used and understood by them. A linguistic sign is made up of a signans representing a signatum determined by convention. Although all linguistic signs are symbols, Peirce emphasises the fact that many of them are also motivated to some extent. In order to show the degree and kind of this motivation he introduces a semiotic definition of an index and an icon. Peirce classifies signs according to the type of relationship existing between signans and signatum, i.e. the extent to which this relationship is natural or arbitrary. Among several categorizations introduced by Peirce, the one that plays a crucial role in my considerations is the triad of the symbol, index and icon, as in the analysis of morphological iconicity in children’s language I will adhere to the theory of Natural Morphology72 based on this particular classification of Peirce’s. According to Peirce, an index is a linguistic sign exhibiting a natural relationship between signans and signatum. It is a sign that directly points to an object, without describing it, in such a way that a listener can infer from it the existence of a given object. According to Peirce, typical indices in language include proper nouns, demonstratives, pronouns, grammatical morphemes and deictic expressions. For example, any proper noun can be considered to be an index in language because it points to a specific object or person. Deictic expressions constitute typical examples of indices, as they indicate a particular time, place, person or object, e.g. then, there or that. This phenomenon is referred to as the rule of indexicality. Thus, it can be said that thanks to the rule of indexicality we possess the ability to point to things that are in the centre of our attention. As far as the icon is concerned, Peirce defines it as a linguistic sign whose signans shows a relation of similarity or analogy with its signatum. In other words, it is a sign similar to the object it denotes. Peirce (1965.II: 158) claims that icons are the most natural signs: The only way of directly communicating an idea is by means of an icon because in icons there is an intrinsic connection 72 A comprehensive presentation of the tenets of Natural Morphology is going to be presented in the section devoted to the existence of motivated signs in the morphological structure of language, i.e. in subchapter 3.2.1.
124
between signans and signatum. Depending on the degree of analogy and similarity between signans and signatum, Peirce distinguishes three types of icons: images, diagrams and metaphors. In his view, images are the most natural icons, whereas metaphors are the least ‘iconic’ icons. An image is an icon which directly represents the features of an object. Extralinguistically speaking, a photograph is a typical image of what it represents. In language, images are mainly onomatopoeic words, e.g. to twitter, because they imitate natural sounds. A diagram is an icon which represents the relations, mainly dyadic, or so regarded, of the parts of one thing by analogous relations in their own parts (Peirce (1965 II: 157)). A verbal diagram is an icon which shows an analogy of structure between signans and signatum. Diagrams are the most important for Natural Morphology, as it seems that all components of language share some fundamental traits of diagrammaticity. By the phenomenon of diagrammaticity it is meant that many morphological signs exhibit a relation of equivalence between signans and signatum (Wurzel (1994b)). For instance, Indo-European languages express the degree of adjectives by a gradual increase in the number of phonemes. Here is the comparison of adjectives in three languages: English: long – longer – the longest Polish: zimny ‘cold’ – zimniejszy ‘colder’ – najzimniejszy ‘the coldest’ Latin: clarus ‘famous’ – clarior ‘more famous’ – clarissimus ‘the most famous’ On the basis of the above examples it is clear that the shape of the signantia reflects the intensity of gradation conveyed by the signata. Lastly, a metaphor is an iconic sign characterised by similarity to its object. A verbal metaphor is an icon exhibiting some kind of parallelism or partial similarity between signans and signatum. All types of morphological conversion, e.g. bottle – to bottle, water – to water are metaphors because they show signantia mapped onto different, though morphologically related, signata. However, in the light of cognitive linguistics, iconicity in language can be observed not only on the morphological but also on the phonological and syntactic level (Dirven (2001)). When it comes to iconicity in phonology, it is discernible in, for example, English in the structure of pro-forms and question words. The characteristic feature of these linguistic items is that pro-forms begin with a consonantal cluster th-, whereas question words begin with another consonantal cluster wh-. Iconicity in these cases results from the fact that the initial sounds of both pro-forms and question words indicate their meaning and function in a sentence. Thus, it is clear that if a given word begins with, for example wh- it will be used to ask questions 125
(what/when/where/who?), and if it begins with th- it will be used to replace either adverbials of time or place, or subjects or objects (then/there/that). Syntactic iconicity should be understood as a parallelism between the syntactic construction and its semantic construal. Syntactic iconicity may concern the rule of sequentiality, quantity and distance. When talking about sequentiality, it is meant that the order of events is reflected in the linear structure of a sentence, e.g. He entered the shop and asked for a bar of chocolate. The event that happens first also comes first in the sentence, i.e. he first entered the shop, then he asked for a bar of chocolate. By the rule of quantity the tendency to intensify the meaning while increasing ‘the amount’ of form is indicated. This rule is applied very often by small children who express the notion of plurality by a multiple repetition of the appropriate word: Look daddy – tree, tree and another tree. Another example of iconicity in a sentence would be the rule: ‘the more polite you want to be, the more you need to say’ (Taylor (2002: 29)). For instance, the request: I wonder if you could possibly give me a lift? sounds much more polite than: Can you give me a lift? as ‘the amount’ of form is much larger in it. Finally, the rule of distance means that those sentence elements which are ‘close’ to each other notionally also occur close to each other in syntactic structures: 1. I made John leave. 2. I believed that John would leave. In the first sentence the subject (I) exercises a direct influence on the object (John) and hence the distance between the two verbs has been reduced to the minimum. In the second sentence the subject is not trying to influence the object in any way and, because of this, the distance between the two verbs is greater. Summing up, despite the overall arbitrary nature of language, numerous elements exhibit characteristics both of indexicality and iconicity. This means that the structure of these elements is not arbitrary, as it exhibits a logical link with their meaning on the phonological, morphological and syntactic level. As the purpose of this chapter is to analyse and describe the issue of morphological and syntactic iconicity in children’s speech, the following sections will be devoted to a discussion of these two linguistic phenomena. The aim of the succeeding subchapters is to prove that iconicity plays a crucial role in the processes of both inflection and word-formation. For this purpose I will present the main assumptions of the theory of Natural Morphology, which regards iconicity as a crucial parameter for the organisation of morphological components, and which uses Peirce’s classification quite extensively to account for several morphological phenomena. 126
3.2. Morphological iconicity 3.2.1. Natural Morphology Before I discuss the standpoint of Natural Morphology on the issue of iconicity, I will present the main assumptions of the theory of Natural Morphology from a more general perspective. Natural Morphology is the morphological theory of Natural Linguistics (Galeas (1998), Dressler, et al. (1987)). As the very term suggests, Natural Linguistics sees the concept of naturalness as its fundamental feature in the same way that Peirce regards naturalness as one of the basic criteria for his classification of signs. Naturalness is established by the proponents of the theory of Natural Morphology on the basis of empirical evidence, i.e. the observation of linguistic items with regard to their distribution, the pace of acquisition by children, susceptibility to language change and language disorders, and the ease of decoding. Mayerthaler, (1981: 2) has drawn the following conclusions from the observations he has made: 1. Not all morphological processes and structures are equally distributed in natural languages. For instance, suffixes are cross-linguistically more frequent than prefixes (Cutler et al. (1985)). Thus, if the language has only one type of affix at its disposal, it will be a suffix (Turkish lacks prefixes but has suffixes). 2. Not all morphological structures are acquired by children at the same time. Irregular, i.e. suppletive paradigms, are acquired later than regular ones in first language acquisition (Chini & Galeas (1995)). 3. Not all morphological structures are equally affected by language change. For example, the operation of suppletion is typologically dependent in this respect (Dressler (1985c, 1986)). Since inflecting languages show more suppletion than agglutinating languages, a language which changes from one type to the other increases or decreases its amount of suppletion accordingly. Estonian, which has become an inflecting language, has seen a rise in the number of suppletive forms, but Tokharian, which has become an agglutinating language, has seen a drop in suppletion. 4. Not all morphological processes and structures are equally impaired by language disorders. Dressler (Dressler & Denes (1988)) has studied the decoding of complex words by aphasiacs and found that they handle more transparent words better than less transparent ones. Transparency should be understood here as morphosemantic compositionality, as in the derivative reader, meant to denote
127
‘someone who reads’, where meaning of the parts: read and -er yields the meaning of the whole, as the first part, i.e. read stands for the activity, whereas the second part, i.e. -er for the doer of this activity. Such words are, according to Dressler, easier to process both in production and perception than the words in which morphosemantic compositionality has not been preserved, as in, for example, the noun boiler, standing for ‘the chicken meant to be cooked’ and characterised by opaque, i.e. non-compositional meaning. 5. Not all morphological structures are equally easy to decode. This means that words not complying with the ‘one meaning – one form’ principle require much more effort in morphological processing. For instance, English uses four suffixes to form similitudinal adjectives, e.g. -like in beast-like, -ly in daughter-ly, -y in chalk-y and -ish in clown-ish. However, these suffixes also represent other meanings. For example, -ly is also used to form adverbs, as in slow-ly; -y forms possessional adjectives, as in dirt-y; and -ish is used to make attenuative adjectives, e.g. pink-ish. Thus, Natural Morphology predicts that these ambiguous suffixes should be decoded with much more difficulty than a biunique suffix, such as, for instance -able, which has only one meaning ‘fit for being V-ed/liable to be V-ed’ (Marchand (1969: 230)). On the basis of these conclusions, naturalists have put forward a list of premises presented below. A morphological process or a morphological structure is natural if it is: a. widely distributed and/or b. acquired relatively early and/or c. relatively resistant to language change or develops frequently by language change and/or d. relatively less likely to be impaired by language disorders and/or e. relatively easy to decode. In other words, naturalness is equated in this approach with unmarkedness, which means that the less marked a given structure is, the more morphologically natural it is. Consequently, a morphological phenomenon is considered to be unnatural if it is marked. The notion of naturalness or unmarkedness is explained by naturalists in relation to the extralinguistic foundations of language, which can be divided into neurobiological (including psychological) bases and socio-communicative (including socio-psychological) bases (Wurzel (1994b)). The set of neurobiological bases includes psychological limitations of perception and receptive processing, limitations of memory, restrictions on storage and retrieval of information, etc. Socio-communicative bases are connected with the communicative function of language, e.g. the adjustment of the speaker’s articulation of the message to its reception by the listener for the sake of ease of 128
perception (Clark (1996)). Thus, human capacities determine which operations are more or less natural. To sum up, what is easier for the potential language user is what is called natural. In other words, cognitive simplicity equals naturalness, i.e. unmarkedness. Galeas (1998) has proposed the existence of eleven parameters of universal naturalness that express the preferences or tendencies in human languages in the choice of morphological operations and rules with respect to the ease of morphological processing, i.e. the degree of their naturalness. These parameters are scalar,73 not binary factors, which means that some morphological operations are more or less natural, as well as more or less unmarked, i.e. more or less easy for the human brain. Thus, the range of options within a single parameter runs from the minimum to the maximum of naturalness. These are the parameters of morphological naturalness:74 1. Diagrammaticity. A word is diagrammatic if it is perfectly segmentable and semantically motivated. A typical example of such a word is sing-er, which can be easily divided into two parts: a verbal base sing and the derivational suffix -er, denoting an agent. In this case morphotactic transparency diagrams (i.e. reflects analogically) semantic compositionality. In contrast, the verb went is neither segmentable, nor semantically motivated, therefore, it is not diagrammatic. 73 The concept of scalarity has been discussed by Galeas (1998) and Tabakowska (1995). Galeas defines it as an intermediate approach to the phenomenon of categorisation, perceived by cognitive linguistics as a process crucial for the organisation of linguistic input. Galeas proposes that his approach mediates between […] the rigidity of the discreetness of the classical principle and the elusive vagueness of the prototypicality principle (Galeas (1998: 25)). By the discreetness of the classical principle is meant the Aristotelian theory of category, which proposes that the categories, once established, are fixed and unchangeable, and their defining features are binary (Galeas (1998)). This makes the whole theory inflexible. In contrast, the elusive vagueness of the prototypicality principle refers to the theory put forward by Eleanor Rosch (1978) who has stated that categories are based on a prototype and are subjective, because they are formed by every language user individually in the process of concept formation. According to Galeas, categories are neither as rigid as those Aristotelian proposed nor as flexible as those proposed by E. Rosch. Therefore, he puts forward scalar categorisation, which proposes that the items to be classified should be placed on a scale (understood as the opposition of extremes) according to the extent to which they exhibit a particular feature. 74 Some of these principles of morphological naturalness are convergent with the operating principles proposed by Slobin (1985) by means of which children perceive their linguistic environment and try to organise it (these principles have been presented in the first chapter). For example, Operating Principle D – underlying semantic relations should be marked overtly and clearly – coincides with parameters 1–6 of morphological naturalness which ensure the clarity of both semantic and grammatical encoding. Operating Principle C – avoid exceptions – is convergent with the overall premise of morphological naturalness according to which marked structures, i.e. exceptions, are acquired late in the course of the language acquisition process.
129
2. Morphotactic transparency. This parameter is defined as boundary recognisability which pertains to the morphemic segmentation of a complex word. For example, kot ‘cat’ – kot-y ‘cats’ are morphotactically transparent complex words as one can easily separate the suffix from the base. On the other hand, suppletive forms, such as: rok ‘year’ – lata ‘years’ are not transparent as they cannot be segmented. 3. Morphosemantic transparency. The semiotic principle underlying the parameter of morphosemantic transparency is the so called Fregean principle of compositionality of meaning (Galeas (1998)) which states that the meaning of a complex word is a function of the meaning of its constitutive parts. Thus, a compound noun teacup is morphosemantically transparent, whereas telephone box is morphosemantically opaque because it is a lexicalised item. 4. Uniformity. A uniform sign is a sign whose signatum is expressed by a single signans. For example, in English the progressive aspect is encoded only by the suffix -ing which is therefore a uniform encoding sign. On the other hand, in Polish 1st person SG can be encoded by several inflectional suffixes, such as: -ę (the first and the second conjugation) in pisz-ę ‘I write’, -am (the third conjugation) in słucham ‘I listen’ or -m (the fourth conjugation) in wiem ‘I know’ (Bąk (1989)). Therefore, it can be claimed that the signatum ‘the first person SG’ is not realised in a uniform manner in Polish. 5. Transparency of encoding. This parameter implies that a given signans represents one and only one signatum. In contrast, opaque encoding should be regarded as a situation in which one signans encodes more than one signatum. Thus, the English inflectional suffix -s does not represent transparency of encoding because it is used to denote the following grammatical morphemes: the third person SG, as in he write-s; plural, as in pen-s; and the genitive form, as in Bob’s. 6. Biuniqueness. This parameter implies that a complex sign is biunique if each part of its signans expresses always and only its corresponding signatum. Biuniqueness is quite rare in world languages. One language in which it can be observed is Modern Greek, in which the prefix is- conveys just one meaning, i.e. ‘movement toward or in a place’. At the same time, this meaning is always rendered by the prefix is- as in pnoi ‘breath’ and is-pnoi ‘inhalation’. 7. Indexicality. This parameter derives from an index and it implies that the relation between an index and its object is that of contiguity, i.e. it measures the capacity of a given sign to refer itself to another sign. The indexical force is a function of proximity between an index and an indexed sign. The most natural indices are derivational morphemes, which are the most direct as they indicate the lexical value of the signatum, e.g. -en in whit-en indicates a deadjectival verb. On the other hand, the inflectional affixes which refer to 130
the grammatical content of the signatum are relatively less natural as they are more peripheral to the lexical morpheme. 8. Metaphoricity. This parameter allows the evaluation of complex signs, i.e. morphological metaphors, characterised by a partial similarity between signans and signatum. Morphological metaphors are semantically complex but morphotactically unanalysable as they do not show the analogy of structure between signans and signatum. In English de-adjectival abstract nouns, such as: the young, the rich, etc. are typical examples of morphometaphoricity. On the other hand, English abstract nouns are less typical examples because they can be either sources of verbalisations: hate – to hate, or targets of nominalisations: to hate – hate, i.e. they lack a good criterion of directionality. 9. Size of the signans. This parameter specifies the parameter of natural size, i.e. length of the signans (lexical, derivational and inflectional morphemes) that would be the most frequent, i.e. the most natural. For example, English monosyllabic derivational morphemes are very natural because of their high frequency. 10. Morphological base. According to this principle, free or bound lexemes are the most natural as morphological bases, which means that they can serve as a basis for the formation of the largest number of derivatives, e.g. use – useful, useless, used, user, misuse, disuse, etc. In contrast, complex words are far less natural bases for morphological rules, e.g. useful – usefully, usefulness. 11. Symbolicity. This parameter derives from the semiotic principle of symbolicity. From a morphological point of view, a symbol is any word that is semantically complex but morphotactically unanalysable as it is bereft of internal structure. Prototypical symbols are inflectional or derivational suppletive words. On the whole, the theory of Natural Morphology proposes a comprehensive list of features that a morphological process should be characterised by if it is to be considered natural, i.e. unmarked. In the light of what has been said above, cognitive and linguistic simplicity equals morphological naturalness. In the following subchapters I am going to discuss the issue of naturalness, i.e. unmarkedness in language, understood as iconicity, within the framework of the theory of Natural Morphology. 3.2.2. Morphological iconicity in the light of the theory of Natural Morphology Natural Morphology regards semiotics as a superordinate framework or metalevel (Galeas (1998: 7)). Therefore, it has adopted the system of signs put forward by semiotics (Peirce (1965)) for the purpose of morphological
131
analysis. Considering the whole array of linguistic signs, Natural Morphology sees an icon as the most natural sign because of the existing analogy between signans and signatum (Dressler (1985)). Of the three iconic signs – image, diagram and metaphor – it is the diagram that is of primary importance for the theory of Natural Morphology because it makes it possible to analyse the structure of words and the organisation of the morphological material with a view to establishing its degree of naturalness or unmarkedeness. In what follows my aim will be to show that the majority of morphological and syntactic operations employed by children in the initial stages of their linguistic development exhibit a high degree of both morphological and syntactic naturalness. What is more, I will prove that the language of children is much more iconic than the language of adults, as children resort to natural morphological and syntactic operations in cases in which adult language depends on symbolic, i.e. marked structures. In other words, my aim is to show that the language of children is much more regular than the language of adults. For the purpose of the analysis of morphological naturalness, I will concentrate on the discussion of all three subtypes of icon: image, diagram, and metaphor. Natural Morphology accounts for the fact that children rely on iconic paradigms much more often than adult users of language by claiming that iconic structures take psychological reality into account (Dressler (1985b)). This means that, according to the theory of Natural Morphology, what is easier from the cognitive and linguistic perspective will be favoured in linguistic production. Therefore, iconic forms are preferred by children to symbolic ones, as they are more overt and, therefore, their structure can be easily analysed. At the same time, morphological rules governing their production are more ‘logical’ and, consequently, simpler to comprehend and follow. For example, first language acquisition research has shown that the process of learning English past tense verb forms is full of overregularisations (Slobin (1985)). This is so because regular verbs when put into past tense are perfectly diagrammatic, i.e. their morphotactic transparency equals semantic compositionality. The addition of the inflectional suffix -ed is accompanied by the addition of meaning: ‘this action was performed in the past’. In contrast, irregular verbs are morphologically unnatural, i.e. their formation is against both the rules of morphotactic transparency and semantic compositionality. Therefore, irregular past tense forms are acquired later: 75 at first children overregularise 75
At this point I mean the stage at which children actively build their language by decoding the rules governing its structure and trying to employ them. In the light of the
132
irregular verbs by constructing such forms as: go-ed, eat-ed; only later do they realise that the past tense can also be formed suppletively. In what follows I will discuss the reasons for the occurrence in children’s language of all the three types of icon – image, diagram and metaphor – and analyse their internal structures. I will juxtapose these children’s iconic forms with symbolic adult structures and, on the basis of the juxtaposition, prove that in the vast majority of cases children adhere to the principle of morphological naturalness when acquiring their native language. First of all, I will concentrate on a discussion of images, which occur quite frequently in children’s speech at the beginning of their linguistic development. Then I will focus on the issue of diagrammaticity in children’s speech in order to examine the whole array of simplifications that children resort to as a result of subscribing to the principle of morphological naturalness. Lastly, I will debate the question of morphological metaphors in the word formation process. 3.2.3. Images in children’s language I will begin the discussion of morphological iconicity in children’s speech by examining the occurrence of images. The following subchapter constitutes an attempt to analyse the morphological structure of images as well as the cognitive mechanisms that lead to their production. To begin with, I will present a reminder of the semiotic definition of an image: Peirce defines an image as a subtype of an icon which directly represents certain features of an object. From a traditional point of view, onomatopoeic words are typical images because they embody the sounds produced by people, animals or objects. As far as first language acquisition is concerned, on the basis of the present data it can be stated that onomatopoeic expressions can be heard in children’s speech quite often. However, they are different from those encountered in adult language, as they occur in the form of reduplicated phrases. My Polish data abounds in numerous examples of this kind: cognitive theories of first language acquisition this stage is preceded by a phase in which children do not possess abstract concepts and are incapable of abstract thinking (Croft (2001), Dąbrowska (2000)). At this stage the only linguistic progress they make consists in memorising separate lexemes or whole phrases without relying on abstract thinking or grammatical knowledge. At this very first stage of language acquisition, when the system of internalised grammar is non-existent, English-speaking children are heard to ‘use’ correctly irregular past tense verb forms not because they have learnt the rule and know how to use it but because they have learnt a few chosen verb forms by rote.
133
K (1;7–1;10) iconic onomatopoeic reduplication tup-tup brum-brum zium-zium puk-puk miau-miau hau-hau ko-ko ciur-ciur na-na kap-kap kum-kum tiku-tiku
symbolic ‘adult’ gloss ‘tupać’, i.e. ‘stamp one’s feet’ used to describe the sound that a car makes, or a car used to describe the sound that a zip makes, or a zip ‘pukać’, i.e. ‘knock’ ‘miauczeć’, i.e. ‘to miaow’ ‘szczekać’, i.e. ‘to bark’ ‘gdakać’, i.e. ‘to cluck’ ‘ciurkać’, i.e. ‘to trickle’ ‘śpiewać’, i.e. ‘to sing’ ‘kapać’, i.e. ‘to drip’ ‘kumkać’, i.e. ‘to croak’ ‘tykać’, i.e. ‘to tick’
The above iconic onomatopoeic expressions are formed in two ways: 1. by reduplication of the initial syllable of a verb which is onomatopoeic as well: tup-tup from tupać ‘to stamp one’s feet’, or miau-miau from miauczeć ‘to miaow’. As far as these onomatopoeic verbs are concerned, it is the first syllable that is onomatopoeic: ciur-, kum-, puk-, miau-. The second syllable merely constitutes the inflectional ending – in this case the infinitive -ać or -eć, sometimes preceded by stem final consonants, such as -k in ‘ciurkać’, ‘to trickle’, or -cz in ‘miauczeć’, ‘to miaow’. In other words, it conveys no distinct meaning from the child’s point of view and is therefore deleted; 2. by coining an onomatopoeic syllable, resembling a given sound: brum, zium, etc. and reduplicating it: brum-brum for a car, or zium-zium for a zip. In both of these cases the children’s utterances are more onomatopoeic, that is more iconic, than adult ones. By claiming that they are more iconic it is meant that they are more natural, as they imitate particular sounds more accurately, namely by reduplicating an onomatopoeic syllable of the verb, disregarding the inflectional ending, or by relying on onomatopoeic coinages, non-existent in adult language. Taylor (2002: 46) refers to this phenomenon as imitative iconicity. As has been shown above, when compared with children’s language, adult language is either not onomatopoeic (e.g. śpiewać, ‘sing’ or gdakać ‘cluck’ when contrasted with na-na or ko-ko) or far less onomatopoeic, as in miauczeć, ‘to miaow’ or ciurkać ‘to trickle’, when contrasted with miau-miau or ciur-ciur. The very process of reduplication enhances iconicity because the repetition of syllables stands for the repetitive nature of the sounds in the real 134
world. Thus, it can be said that although reduplicative onomatopoeic expressions are examples of images, they carry traces of diagrammaticity as well, as the repetition of syllables diagrams, i.e. reflects analogically, the repetition of the sounds heard in the world we live in. All the examples I have analysed so far come from the Polish data, as in English the mechanism of reduplicative onomatopoeic expressions is not so widespread as in Polish (Szymanek (1998: 75)). 76 Consequently, I have recorded no instances of onomatopoeic reduplications in the English data. As far as the literature and the study of the acquisition of English is concerned, the issue of reduplicative phrases has been ignored; therefore, for the discussion of reduplicative images I will concentrate exclusively on the Polish data. 77 Onomatopoeic expressions do not constitute the only kind of images that the present corpus contains. Here I would like to advance the thesis that in children’s speech there also occur images that are not onomatopoeic, e.g. piszu-piszu. I would like to venture a statement that the phrase piszu-piszu can be labelled an image, as this phrase is a direct representation of the feature of the object it stands for. Yet, in this case, it is not the sound that this object produces but about the activity denoting its inherent and primary function. Thus, piszu-piszu, from pisać, i.e. ‘to write’, stands for a ballpoint, whose primary and inherent function is that of writing. Other examples of this kind are the following: iconic reduplication czeszu-czeszu78 kic-kic
symbolic ‘adult’ gloss ‘grzebień’, i.e. ‘comb’ from czesać się, i.e. ‘to comb’ ‘królik’, i.e. ‘rabbit’ from kicać, i.e. ‘to jump’
76
The phenomenon of reduplication in English tends to be discussed more in the context of compound-formation than child language (Jespersen (1965), Marchand (1969)). 77 The issue of reduplication in child Polish has been discussed by, for example, DziubalskaKołaczyk (1997). 78 The characteristic feature of many iconic reduplicative phrases denoting actions is their -u ending. Dziubalska-Kołaczyk (1997: 167) has remarked the consistent use of -u in the vowel final position in the case of both nouns and verbs. I believe that the reason for the overuse of -u ending can be sought in the language input aimed at children which comprises numerous nouns and verbs containing the vowel -u in their final syllable. These include: 1. masculine diminutives: kotuś ‘kitten’, dziadziuś ‘grandpa’; 2. the genitive and dative of some masculine nouns: płotu ‘fence’ (gen.), kotu ‘cat’ (dat.); 3. imperative verb forms: narysuj ‘draw’, ugotuj ‘cook’, etc. Since vowels are more perceptually salient than consonants, children probably perceive -u as a word final sound, disregarding the consonant that follows it. In addition to this, diminutives and imperatives are quite frequent in caretaker speech; consequently, at the onset of the verbal stage children consider the -u ending to be more frequent than it really is and, as a result, overgeneralise it, by, for example, making it a final sound in the reduplicative phrases.
135
mieszu-mieszu opa-opa usiu-usiu tup-tup
‘łyżka’, i.e. ‘spoon’ from mieszać, i.e. ‘to mix’ ‘piłka’, i.e. ‘ball’ from robić op/opa,79 i.e. ‘to hop’ ‘huśtawka’, i.e. ‘swing’ from huśtać się, i.e. ‘to swing’ ‘schody’, i.e. ‘stairs’ from tupać, i.e. ‘thump’
The above utterances can be regarded as images because the reduplicative phrases they are made of stand for particular actions, representing specific objects, e.g. kic-kic stands for ‘królik’, i.e. ‘rabbit’. In other words, reduplicative phrases imitate by the structure of their signans the characteristic features of their signatum. A discerning reader may voice reservations about the reliability of this classification. Whereas it seems fairly unquestionable that writing is the basic function of a ballpoint, and that combing is inherently associated with a comb, it does not seem so indisputable that the basic feature of a spoon is its capability of mixing or stirring things, or that jumping is the most characteristic feature of a rabbit. Perhaps the basic function of the spoon is to help us eat and the most salient feature of the rabbit is that it has long ears? In fact, it is impossible to give an unequivocal answer to this question, as in the light of cognitive theories of language acquisition, the categorisation of both linguistic and extralinguistic reality is entirely subjective, which means that every human being concentrates on different aspects of reality (Dąbrowska (2003)). In order to explain the phenomenon of subjectivity in the categorisation process I would like to resort to Langacker’s concept of the viewing metaphor, whose point of reference is visual perception (Langacker (1987)). This means that the viewer has in view the so called maximal field of view, i.e. the area that he can theoretically take in. However, at any particular moment of the observation process this field is narrowed down to the area that the viewer’s attention is concentrated on. This smaller field is referred to as the viewing frame. In an attempt to define it, Langacker uses a theatrical metaphor. Namely, he claims that the viewing frame constitutes that part of the stage that attracts the viewer’s attention. Next, he relates visual perception to mental experience, stating that the viewing frame constitutes what I focus on at a particular moment in our perception of reality. In this connection the objection concerning the unreliability of the above classification can be easily dismissed. According to Langacker, every language user is entitled to possess a different viewing frame, i.e. to concentrate his attention on a different aspect of reality. Thus, for one child it is long ears that constitute the viewing frame of a rabbit, whereas for some other child the 79
This is an expression which is non-existent in adult speech; in child language it means ‘skakać w górę’, i.e. ‘to hop’.
136
viewing frame of a rabbit is its activity of jumping and for yet another it is its strong teeth, etc. Hence, a multitude of viewing frames and possibilities exist within the categorisation process. On the whole, this kind of image is non-existent in adult language, which uses conventional, i.e. arbitrary forms. As in the case of onomatopoeic expressions, reduplication reflects analogically the repetition of an action; czeszu-czeszu refers to the action of combing that is repeated. Summing up, an image is an iconic sign which finds a much wider application in child speech than in adult language thanks to its cognitive and linguistic simplicity and high degree of naturalness, resulting form the fact that it represents the features of an object in a straightforward way. 3.2.4. Inflectional diagrams in children’s language As has been shown above, diagrammaticity constitutes one of the parameters of universal naturalness, and, like all the other parameters, it is considered to be a scalar factor. This is because one of the tenets of Natural Morphology is the theory of scalar categorisation, which proposes that a given set of linguistic items is always placed along a continuum, measuring a feature in question. Consequently, all the morphological units are placed along a scale of diagrammaticity, ranging form the minimum to the maximum value and, on this basis the degree of their diagrammaticity is determined. As a result of this, all words in any language can be classified according to the degree of diagrammaticity they exhibit. Thus, some lexemes are considered to be perfectly diagrammatic, e.g. paint-ed, where the morphotactic transparency diagrams, i.e. reflects analogically, semantic compositionality. Others are regarded as not diagrammatic, e.g. go - went, because of the lack of both morphemic segmentation and semantic compositionality. However, the majority of forms exhibit an intermediate degree of diagrammaticity, e.g. shelves, where morphosemantic transparency exists but is disturbed by the modification of the base signans. Furthermore, according to the assumptions of Natural Morphology, those morphological operations or processes that are natural are acquired relatively early in the process of the first language acquisition. Consequently, the following conclusion can be drawn: small children acquiring their native language will rely much more often on the process of agglutinative affixation or affixation + modification, which are considered to be the most natural, than on modification, metaphoricity, total suppletion or subtraction, which are considered to be quite unnatural. This statement has been verified on the basis of the Polish and English data and it has turned out that natural structures are given priority by children in the process of first language acquisition. 137
On the whole, the diagram is of primary importance for the theory of Natural Morphology because although its degree of iconicity is intermediate, i.e. it lies between that of an image and a metaphor, it constitutes an indispensable tool for examining the internal structure of words. In the analysis of the children’s speech I will make use of the scale of diagrammaticity proposed by Galeas (1998) in order to establish the extent to which the language of children is diagrammatic. The scale of diagrammaticity The scale of diagrammaticity comprises six degrees and some subdegrees. Its aim is to establish the degree to which there exists a parallelism between the sequence of morphemes and the composition of meanings. 1. First degree: Agglutinative affixation. This represents the most natural degree of the scale of diagrammaticity because it consists in the direct affixation onto the lexical base. Examples of this process would be some English prefixed verbs, like re-write, mis-use, and un-tie, and some Nomina Actionis, e.g. conceal-ment, or adapt-ation. As far as Polish is concerned, examples of diagrammatic words would be some plural nouns, such as dom-y ‘houses’, samolot-y ‘planes’, or komputer-y80 ‘computers’. These words are thoroughly diagrammatic, as they are both morphotactically transparent and semantically motivated. 2. Second degree: Affixation + modification. The second degree of the scale is characterised by a lower degree of diagrammaticity because the process of affixation entails a modification of the lexical morpheme, e.g. shelf – shelv-es, or study – stud-ies, where morphonological rules affect the final sounds of the two lexemes. Diagrammaticity is preserved due to the operation of suffixation, whereas morphotactic transparency is disturbed by the modification of the base signans. 3. Third degree: Modification. This degree is even less natural than the preceding two, as it consists in the mere modification of the lexical items without affixation, so, as a result, the amount of iconicity is minor. The words that are characterised by the third degree of diagrammaticity are, for example, English plural irregular nouns: foot – feet, goose – geese, tooth – teeth, etc. The same phenomenon can be observed in, for example, German, as in Vater – Väter. 80
Note that the inflectional ending -y constitutes an instance of the so called cumulative exponence, i.e. the phenomenon characterized by the simultaneous realization of several grammatical features, in this particular case: number (plural), case (nominative) and gender (masculine).
138
4. Fourth degree: Metaphoricity. This degree is characterised by the absence of any operation of affixation or modification, i.e. the inalterability of a base which is a constant feature of all metaphorical rules. The output of a morphological metaphor is the same as the input. Thus, the morphotactic level does not reflect the semantic modification. For example, conversion represents a typical example of metaphoricity, as in: a drink – to drink, a book – to book, etc. 5. Fifth degree: Total suppletion. This represents an anti-iconic operation showing no diagrammaticity, i.e. no morphemic segmentation and no semantic compositionality. Total suppletion is cross-linguistically rare and unproductive, and limited only to a few morpholexical domains, e.g. irregular comparison of adjectives both in English and in Polish: dobry ‘good’ – lepszy ‘better’, or zły ‘bad’ – gorszy ‘worse’. 6. Sixth degree: Subtraction. This represents an anti-iconic operation, as the deletion of the phonological material means the addition of meaning, which goes against one of the basic principles of iconicity, according to which the addition of phonological material entails a change in meaning. Because of this, anti-diagrammatic subtraction is the least natural morphological operation as well as the least productive. It can be observed in just some languages in a few morpholexical domains, e.g. in Polish animate nouns denoting people engaged in scientific disciplines: fizyka ‘physics’ – fizyk ‘physicist’, matematyka ‘mathematics’ – matematyk, ‘mathematician’, geografia ‘geography’ – geograf ‘geographer’, etc. As can be seen, the intensity of diagrammaticity decreases with each degree: agglutinative affixation turns out to be the most diagrammatic, whereas subtraction is the least diagrammatic morphological operation. Diagrammaticity equals naturalness, i.e. unmarkedness, which means that the more diagrammatic a given morphological operation is, the more natural or unmarked it is. Thus, agglutinative affixation is considered to be unmarked, whereas subtraction is highly marked and unnatural. According to the assumptions of the theory of Natural Morphology (Dressler (1985)), the more diagrammatic a given structure is, the more natural, i.e. easier to decode, it is and, therefore, the faster it is acquired by children. The basic reason for this is that Natural Morphology takes psychological reality into account, claiming that what is easier for the potential language user is learnt earlier. As there is no doubt that regular grammatical rules are easier to decode, comprehend and follow than irregular ones, morphological processes which do not involve suppletion or which are suppletive only to a negligible extent are the first to be mastered by the youngest users of the language. These processes are: agglutinative affixation 139
and affixation + modification. For example, the inflectional Present Progressive morpheme -ing is among the first to be acquired by Englishspeaking children (Brown (1973)) due to the fact that its addition does not disturb the morphosemantic transparency of the verb in any way. Another aspect of this phenomenon is that children not only learn regular rules before irregular ones, they also resort to regular forms in cases in which irregular structures are required. Such forms created by children are unacceptable in adult language, which in such cases employs the rules of suppletion or modification. Children in the initial stages of their linguistic development do not obey the rule of suppletion, overregularising many grammatical forms and making them more morphologically transparent and semantically compositional than their adult counterparts. For instance, English-speaking children overregularise the plural and past tense formation, creating such forms as foots, or taked. Therefore, on the basis of the analysis of the data, I would like to advance the thesis that the language of children is much more diagrammatic, i.e. regular, that the language of adults on the level of both inflectional and derivational morphology. In what follows, I will discuss the phenomena of morphological overregularisation and simplification in the language of children, showing that they make it far more natural, i.e. less symbolic, than the language of adults. To this end, I will analyse my corpus in the following way: 1) First of all, I will examine the way in which Polish-speaking children simplify and overgeneralise the Polish inflectional system; 2) Secondly, as the inflectional system is in English virtually non-existent, I will concentrate on the rare cases of simplifying grammatical structures by English-speaking children; 3) Thirdly, I will prove that diagrammaticity constitutes one of the possible manifestations of the principle of iconicity in the word formation processes both in Polish and in English. In the next subchapter I will discuss the way in which Polish children stick to the principle of morphological naturalness in their language, making it for this reason far more diagrammatic than the language of adults. 3.2.4.1. Inflectional diagrammaticity in Polish 81
As has already been stated, diagrammaticity constitutes one of the parameters of Natural Morphology. Like all the parameters, it is scalar in nature, 81 A substantial parts of this subchapter have been published in Word Structure 4.1 in the article entitled Inflectional Diagrammaticity in Polish: A Case Study of Polish Children (pp. 20–53).
140
which means that particular words or structures are marked by various degrees of diagrammaticity, beginning with words that are completely diagrammatic, like dom ‘house’ – dom-y ‘houses’, through words of intermediate diagrammaticity, like ręka ‘hand’ – ręce ‘hands’, and ending with those that are non-diagrammatic, e.g. rok ‘year’ – lata ‘years’. In Polish, most inflected words are characterised by intermediate diagrammaticity because the phenomenon of alternation, which disturbs morphosemantic transparency, is an inherent feature of the Polish inflectional system. Polański (1971) states that alternation is the exchange of phonemes within a morpheme which is regular in character, i.e. it takes place according to a constant and definite schema. Szymanek (1998: 19) differentiates three kinds of alternation depending on the kind of conditioning that takes place during this process, as alternation may be phonologically, morphologically or lexically conditioned.82 Thus, adult Polish, whose conjugational and declensional paradigms abound in alternation, is not natural, i.e. morphologically transparent and semantically compositional, for young children whose cognitive set-up is constantly searching for explicit rules. Because alternation is an irregular linguistic phenomenon and its acquisition therefore surpasses the cognitive abilities of children in the initial stages of language acquisition, it is ignored by them. Since one of the principles of Natural Morphology states that what is easier for the potential language user will be acquired earlier, children acquire simplified inflectional paradigms, ignoring the existence of alternation, which poses serious cognitive and linguistic difficulties for them. Therefore, child language is more regular than the language of adults. In order to support this statement let me present some examples of morphologically natural child language from the corpus: 82 By a phonologically conditioned alternation Szymanek means the process in which the appearance of allomorphs is conditioned by general phonological rules. For example, in Polish, on the basis of the process of diminutive formation, I observe a phonologically conditioned alternation: but – buc-ik, resulting from the operation of palatalisation and high vowel adjustment (Gussmann (1980b: 87ff)). An example of a morphologically (grammatically) conditioned alternation is the process of pluralization in the course of which the majority of feminine and masculine nouns are marked with the -i or -y ending, e.g. dziewczynk-i ‘(little) girls’, or motor-y ‘motorbikes’. However, there are many other inflectional suffixes which denote the plural. These are: -e, as in koni-e ‘horses’ (masc.), noc-e ‘nights’ (fem.), -a, as in serc-a ‘hearts’ (neuter), or owie, as in pan-owie ‘sirs’ (masc. pers.). Since the distribution of these endings depends on the gender of the nouns, and even on some stylistic or semantic factors, alternation of this kind is considered to be morphologically conditioned. Lastly, the alternation may be lexically conditioned as in the dziecko ‘child’ – dzieci ‘children’ pair, where the appearance of the allomorph [c’] in dziec-i is neither phonologically nor morphologically conditioned but is merely lexeme specific, i.e. can be seen only in the form of this single root.
141
M (2;0) napompowa-j wstawa-my
from napompowa-ć from wstawa-ć
nie prasowa-j dawa-ø bez koł-ów
from prasowa-ć from dawa-ć from koł-a
instead of napomp-uj, i.e. ‘pump up’ instead of wsta-jemy, i.e. ‘we are getting up’ instead of nie pras-uj, i.e. ‘do not iron’ instead of da-je, i.e. ‘she gives’ instead of bez kół, i.e. ‘without wheels’
As can be observed on the basis of the above language sample, the degree of internal modification in child language is minor when compared with the language of adults, in which the most common way of inflecting words proceeds according to a rule which constitutes the second degree of diagrammaticity, i.e. affixation + modification (Bąk (1989: 278)). Thus, for example, wstawa-my is more diagrammatic than wsta-jemy, as it is formed by affixation alone. Wsta-jemy is made by affixation + modification of the stem.83 The present section is devoted to a detailed discussion on the strategies adopted by children that are aimed at making their language more natural, which involves ignoring the phenomenon of alternation. This operation is particularly conspicuous in the process of inflecting nouns and verbs, the basic speech categories acquired and used by children in the initial stages of their linguistic development. In what follows, I will analyse several inflectional processes in the course of which children come up with forms definitely more regular than those heard in the language of adults, as there is either no alternation in them or the alternation has been reduced considerably. These processes involve: I. conjugating verbs with little or no alternation within the infinitive stem; II. inflecting nouns with little or no alternation within the nominative stem; III. pluralizing nouns with little or no alternation within the nominative stem; IV. producing the nominative form of the noun from other case forms; V. comparison of adjectives and adverbs with little or no alternation of the base form; VI. forming momentary verbs. 83
When talking about modification within the Polish inflectional system, I mean modification of the stem not modification of the whole lexeme. The reason for this is that ven basic word forms end with specific inflectional morphemes. Thus, in the process of inflection, speakers of Polish do not add inflectional morphemes but exchange them. For example, when conjugating the verb czyta-ć ‘read’, I can exchange the infinitive ending -ć into the first person SG ending -m, as in czytam ‘I read’, or the second person singular ending -sz, as in czyta-sz ‘you read’, etc. Therefore, when seen in this light, every lexeme is modified in the process of inflection, whereas its stem may or may not remain unchanged.
142
The above-mentioned inflectional processes that have been chosen for the analysis have been ordered according to: 1. the frequency of their occurrence in my data; 2. the degree to which adult and child language differ with respect to their diagrammaticity in using a given structure. Thus, inflectional processes I, II, III, and IV are both common in the data and much more diagrammatic in child language than in adult language due to the fact that they involve a considerable amount of alternation within the word stem in the adult language. Inflectional processes V and VI are less frequent in the data and more regular in the adult language than those from the first group. Consequently, the difference between the degree of their diagrammaticity in child language and in adult language is not as striking as in the previous group, although still conspicuous. The reason for this is that the greater degree of diagrammaticity in the adult language is tantamount to the lesser need for oversimplification in the language of children. The next section is devoted to a morphological process that is extremely frequent in children’s speech and much more diagrammatic than in adult language, i.e. conjugating verbs. Diagrammatic verbs The process of verb conjugation is particularly common in the language of children because, on the basis of the corpus under study it can be stated that it is actions, and things or people that young children most often talk about; hence the high frequency of verbs in children’s speech. The Polish conjugational system is characterised by a significant degree of sophistication and richness of verb forms, especially in the present tense of the indicative mood (Bąk (1989: 322)). My data clearly shows that the indicative mood of the present tense is a form that young children very often resort to, as they tend to devote a lot of their attention to commenting on actions performed by themselves or by other people. However, apart from this, diagrammatic past and future verb forms are common in my data as well. Nonetheless, they are less diagrammatic than present tense verb forms. Moreover, Polish children coin morphologically natural forms of imperative and momentary verbs. What is common for all the below presented diagrammatic verb forms is that they are all derived from the infinitive form. The reason for this is the fact that child language is prototype dependent, which means that the forms that children most often resort to are prototypical, i.e. basic.84 Therefore, it is the stem of the most prototypical verb form, i.e. the infinitive that children stick to for the formation of diagrammatic verbs. 84
Cf. Taylor (1989), Marchman et al. (1991), Meints et al. (1999).
143
I am going to begin my analysis with the discussion of the diagrammatic present tense. Diagrammatic present tense verb forms children’s diagrammatic language K (1;9–2;3); Z (1;8– 2;0))
adult conventional language
Tata wstawa-ø Spa-my Kopa-my Trzepa-m Pra-m Pra-sz? Liza-ø Wyjech-uje Odleca-ją Nie dawa-m Maza-ø Załaż-am Dzioba-ją Skuba-ø
‘Tata wsta-je’, i.e. ‘Daddy is getting up’ ‘Śpi-my’, i.e. ‘We are sleeping’ ‘Kopie-my’, i.e. ‘We are digging’ ‘Trzepi-ę’, i.e. ‘I am beating’ ‘Pior-ę’, i.e. ‘I am washing’ ‘Pierzesz?’, i.e. ‘Are you washing?’ ‘Liż-e’, i.e. ‘He is licking’ ‘Wyjeżdż-a’, i.e. ‘He is leaving’ ‘Odlat-ują’, i.e. ‘They are flying away’ ‘Nie daj-ę’, i.e. ‘I am not giving’ ‘Maż-e’, i.e. ‘He is rubbing out’ ‘Zakład-am’, i.e. ‘I am putting on’ ‘Dziobi-ą’, i.e. ‘They are pecking’ ‘Skubi-e’, i.e. ‘She is pinching’
from wstawa-ć from spa-ć from kopa-ć from trzepa-ć from pra-ć from pra-ć from liza-ć from wyjech-ać from odleci-eć from dawa-ć from maza-ć from założ-yć from dzioba-ć from skuba-ć
When it comes to the present tense verb forms created by children, it can definitely be stated that they are definitely more diagrammatic than those encountered in adult language. However, the degree of their diagrammaticity varies, as they are either completely diagrammatic, like wstawa-ø from wstawa-ć, i.e. ‘(Daddy) is getting up’, or diagrammatic only to a limited extent because they are not completely deprived of alternation, e.g. Załaż-am from założ-yć, i.e. ‘I am putting on’.85 Nonetheless, the alternation that can be observed in the speech of Polish children is not as extensive as that occurring in adult language, where apart from vowel alternation consonant alternation can be observed as well, e.g.: Śpi-my Nie daj-ę
from spa-ć, i.e. ‘We are sleeping’; from dawa-ć, i.e. ‘I am not giving’;
alternation: a-i, s-ś alternation: a-ę, w-j, etc.
In the process of avoiding alternation and, at the same time, making the present verb forms as regular as possible, children overgeneralise the rules of 85 According to Grzegorczykowa (1998: 116), the o-a vowel alternation occurs in the process of aspectual derivation, i.e. in the course of forming imperfective stems from perfective stems, e.g. włożyć – wkładać ‘put in’.
144
verb inflection. They narrow down the scope of all the inflectional paradigms to conjugation III, the easiest one to acquire from a morphological point of view. Conjugation III, which is characterised by a stem ending in -a, for all persons, with no alternation apart from the third person plural in which the stem ends in -j, seems to be the most morphotactically transparent for Polishspeaking children and, therefore, the most natural. 86 Thus, because of the great simplicity of this pattern, the majority of verbs are inflected by children according to the conjugation III inflectional paradigm, although they belong to conjugation patterns I or II, e.g. skuba-m instead of maza- ø instead of spa-my instead of
skubi-ę, i.e. ‘I am pinching’ maż-e, i.e. ‘She is rubbing out’ śp-imy, i.e. ‘We are sleeping’
(conjugation I) (conjugation I) (conjugation II)
The only exception in this process of overgeneralisation, as far as my data is concerned, is the form wyjech-uje ‘is leaving’ which is not inflected according to conjugation III. This form can be classified as an example of conjugation I in which verbs in the third person singular possess the -e ending. However, in this case, confusing the conjugation patterns is not the only divergence from the adult pattern. Another thing is that the form wyjech-uje is made from the infinitive of the perfective verb wyjechać instead of the imperfective one wyjeżdżać ‘to leave’. In my language corpus there are two more examples of verbs whose forms have been made from the perfective verb stems. These are załaż-am ‘I am putting on’ from założyć (perf.) instead of ‘zakładać’ (imperf.) ‘to put on’ and odleca-ją ‘They are flying away’ from odlecieć (perf.) instead of ‘odlatywać’ (imperf.) ‘to fly away’. The use of the stems of perfective verbs for the formation of present verb forms can be accounted for on the basis of research conducted by Brenstiern-Pfanhauser (1930), who discovered that the child tends to build all verb forms from the stem that he learned first. Thus, here it is the perfective verb forms that children learned first. The above described process of overgeneralisation suggests that children stick firmly to the principle of uniformity, according to which a given signatum is expressed by single signans. Thus, in the above cases, the 86 In order to show this let us present the conjugation III inflectional paradigm of the verb czyta-ć ‘read’:
Singular 1. czyta-m 2. czyta-sz 3. czyta-ø
Plural czyta-my czyta-cie czytaj-ą
145
following signata can be easily differentiated: the first person singular, the third person singular and the first person plural as well as their corresponding signantia: -am, -ø and -amy. Thus, the process of overregularisation of the present tense form of the verbs comprises two phenomena: 1. simplifying the conjugation pattern by accommodating all the verb forms to the conjugation III paradigm; 2. ignoring or reducing the principle of alternation in the inflectional paradigm. The next subchapter will focus on a discussion of another iconic paradigm, namely the conjugation of past verb forms. Diagrammatic past tense verb forms As far as my data is concerned, the number of past tense verb forms that are much more diagrammatic in child language than in adult language is negligible. In fact, my data contains just two such examples: children’s diagrammatic language K (1;9–2;3); Z (1;8–2;0))
adult conventional language
Ptaszek odlat-ø
from odlat-ywać
Wytrze-łam
from wytrze-ć
‘Ptaszek odleci-ał’, i.e. ‘The bird has flown away’ ‘Wytarł-am’, i.e. ‘I wiped’
The reason for the fact that there is hardly any difference between adult and child language with regard to the formation of the past tense is that the past tense is morphologically simpler than the present tense, as it is formed by subtracting the infinitive suffix -ć and adding the stem-forming suffix -ł, to which the inflectional endings denoting gender and number are attached (Grzegorczykowa (1998: 256)), e.g. spa-ć ‘to sleep’ – spa-ł-am ‘I slept’, jecha-ć ‘to go’ – jecha-ł-am, ‘I went’, etc. Also, the grammatical rules governing the formation of the past tense are relatively diagrammatic, i.e. uniform. This means that all verbs, irrespective of the conjugation they belong to, take the same endings, e.g. the first person singular takes the following endings: -em (masculine), -am (feminine), and -o (neuter), and for the third person singular there are: -ø (masculine), -a (feminine), and -o (neuter).87 Therefore, the whole pattern is natural and consistent for children. Because of this, they are not forced to rely on the process of overregularisation to simplify its acquisition. 87
A comprehensive set of patterns of inflectional paradigms of the past tense can be found in Grzegorczykowa (1998: 259).
146
The only cases in which children rely on their own strategies involve those past verb forms which are not morphotactically transparent, i.e. ones in which alternation is present. Thus, those past verb forms from my data that are not diagrammatic in the adult language have been subjected to the following processes of overregularisation: 1. forming the past form of the verb odlecie-ć ‘fly away’, i.e. odlecia-ł ‘it flew away’ involves e-a alternation, hence the form odlat which entails no internal modification. What is more, this form has been coined from the imperfective infinitive odlat-ywać by subtracting what the child took to be the infinitive ending, i.e. the suffix -ywać. Smoczyńska (1985) states that the imperfective form is morphologically simpler, so the perfective form is often derived from it – as is the case here – especially as it allows the avoidance of the phenomenon of alternation; 2. forming the past form of the verb wytrz-e-ć ‘wipe’, i.e. wytar-ł-am ‘I wiped’, poses considerable difficulties for children because of the extended stem (Grzegorczykowa (1998: 235)), which occurs only in the infinitive form.88 Apart from this, in the adult pattern there is an rz – r alternation. In contrast, the form created by children, i.e. wytrze-ł-am, is perfectly diagrammatic because its construction follows the rules of the regular past construction, i.e. adding to the infinitive stem the stem forming suffix -ł and the first person singular ending -am. All in all, diagrammatic past verb forms are not a representative example of the phenomenon of the greater iconicity of children’s speech when compared with adult language. The reason is the relative simplicity and regularity of the past tense conjugational paradigm. Diagrammatic future tense verb forms Diagrammatic future verb forms are less frequent in my language sample than diagrammatic present verb forms but more frequent than diagrammatic past verb forms: children’s diagrammatic language K (1;9–2;3); Z (1;8–2;0)) Przebra-m Ja wytrzę-ø Rozebrę się Wysypa-my
from przebra-ć się from wytrze-ć from rozebra-ć się from wysypa-ć
adult conventional language ‘Przebio-rę się’, i.e. ‘I will change’ ‘Wyt-rę’, i.e. ‘I will wipe’ ‘Rozbio-rę się’, i.e. ‘ I will undress’ ‘Wysypi-emy’, i.e. ‘We will empty out’
88
Grzegorczykowa (1998) means by the extended stem the inclusion of -e in the infinitive form wytrz-e-ć ‘wipe’.
147
There are two reasons for the relatively low frequency of diagrammatic future verb forms. First of all, future verb forms appeared in my language corpus relatively late, i.e. when the children studied had began to embrace the concept of the future and express it in their speech, at around the age of 2;6. Also, the formation of the future in adult language involves hardly any suppletion. This is because the future tense of imperfective verbs is formed analytically by using the future form of the auxiliary być ‘to be’ followed by the infinitive form, e.g. będzie pisać ‘(he/she) will write’, or the past form, e.g. będzie pisała ‘(she) will write’. The future tense of perfective verbs is also morphologically simple, as it is formed synthetically by prefixing the present form of the verb: na-pisze ‘(he/she) will have written’. However, there are some verbs which require internal modification when producing their future forms (Bąk (1989)) and their formation involves a considerable degree of simplification in the child language, which has been clearly shown in my data. Thus, as can be seen above, such verbs are much more diagrammatic in child language than in adult language. The forms przebra-m89 ‘I will change’ and wysypa-my ‘we will empty out’ are perfectly diagrammatic as their formation involves no alternation of the infinitive stem. The formation of the verbs wytrz-ę ‘wytr-ę’, i.e. ‘I will wipe’ and rozebr-ę się ‘I will undress’ is also diagrammatic, though less so, as their morphotactic transparency has been disturbed due to the occurrence of alternation (e: ę and a: ę, respectively) after the removal of the infinitive-forming suffix -ć. On the whole, the formation of the future tense is more regular than the formation of the present tense and slightly less regular than the formation of the past tense. Another thing is that diagrammatic future verb forms are very few in number in the child language due to the simplicity of the future tense formation system in the adult language. Therefore, the statement that diagrammatic future forms are characterised by a higher degree of iconicity in child speech refers only to the limited number of forms that are irregular in adult language. Diagrammatic imperative verb forms The imperative is quite common in child language, probably due to its frequent occurrence in the language input, i.e. in the caretaker speech: children’s diagrammatic language adult conventional language K (1;9–2;3); Z (1;8–2;0)) Urwa-j from urwa-ć ‘Urw-ij’, i.e. ‘Tear off’ Skak-ø from skaka-ć ‘Skacz—ø’, i.e. ‘Jump’ 89
The inflectional ending -am is used to form the future tense of perfective verbs belonging to conjugation III.
148
Rozebr-ø lalę Rysowa-j tą Narysowa-j
from rozebra-ć from rysowa-ć from narysowa-ć
‘Rozbie-rz lalę’, i.e. ‘Undress the doll’ ‘Rysu-j tę’, i.e. ‘Draw this one’ ‘Narysu-j’, i.e. ‘Draw’
As can be seen, its formation is much more diagrammatic in child language because in adult Polish the imperative construction is highly irregular, as it entails applying a multitude of inflections and various kinds of alternation.90 However, as the above data shows, children overregularise the whole system by relying on two strategies. First and foremost, they add the inflection -j to the last vowel of the infinitive stem or the stem-forming suffix -a without any internal modification. As in the case of forming the present tense, children have limited their choice of inflectional rules to the one which is morphologically the most transparent. This means that they have chosen a way of forming the imperative which is characteristic of verbs belonging to conjugation III because of its morphotactic transparency and semantic compositionality. An example of forming the imperative of a conjugation III verb is gra-j from gra-ć, i.e. ‘play’. The simplicity of perception and production of this imperative form is evident because of the fact that the morpheme boundaries are easily recognisable and the change of meaning caused by the addition of suffix -j is overt and explicit. Therefore, children prefer this extremely regular pattern of imperative formation to other ones which involve internal modification or even suppletion. Hence the forms Urwa-j instead of Urwij, i.e. ‘Tear off’, or Rysowa-j tą instead of Rysu-j tę, i.e. ‘Draw this one’. Another way of forming the imperative that children resort to is anti-iconic, as it consists in employing subtraction, i.e. subtracting the stem final vowel -a and the infinitive ending -ć, which is also the case in adult language. However, although anti-iconic, children’s imperatives are still more iconic than those encountered in adult language as they do not involve alternation, which can be observed on the basis of the following examples: skak-ø instead of skacz, i.e. ‘jump’ from skak-ać and rozebr-ø instead of rozbierz, i.e. ‘undress’ from rozebrać. To sum up, the formation of the imperative constitutes a representative case of diagrammaticity in the Polish conjugational system. The last case of diagrammatic verb forms that will be discussed is the formation of momentary verbs. Forming momentary verbs Momentary verbs denote an activity, usually a completed one, lasting just one moment (Polański (1971)). They are not very frequent in child language but their iconicity is clearly discernible, as they are formed by two infixes: -nę- and -ną-. 90
The formation of the imperative in Polish has been discussed in Bąk (1989: 353).
149
children’s infinitive adult conventional language diagrammatic language form Zryw-ną-ł zryw-ać ‘Zerwał’, i.e. ‘he plucked’ Trzeba to wylew-ną-ć wylew-ać ‘Trzeba to wylać’, i.e. ‘This should be poured out’ Chcę otwar-ną-ć otwierać ‘Chcę otworzyć’, i.e. ‘I want to open’ Ut-nę-łaś ucinać ‘Ucięłaś’, i.e. ‘You cut’ Wezm-nę-łam wziąć ‘Wzięłam’, i.e. ‘I have taken’ Mogę to wezm-ną-ć? wziąć ‘Mogę to wziąć?’, i.e. ‘Can I take it?’ Potrafię to odep-ną-ć odpinać ‘Potrafię to odpiąć’, i.e. ‘I can unfasten it’ In this case diagrammaticity consists not so much in adhering to the prototypical, i.e. basic, form of the verb but in employing two morphosemantically transparent infixes, -nę- and -ną-, to denote the momentary aspect of the action. The formation of the majority of momentary verbs by children involves internal modification as well but despite this these verbs exemplify the second degree of diagrammaticity, which is affixation + internal modification. Adult forms are less natural in this respect, as the momentary aspect of the action is not marked overtly and explicitly, and can only be inferred from the context, e.g. On szybko wziął dziecko na ręce ‘He picked up the child quickly’. Thus, all in all, children’s forms are still more diagrammatic despite the substantial degree of alternation, as the addition of the infix -nę- or -ną- diagrams directly the semantic change from the infinitive form to the momentary aspect, whose explicit morphological representation is absent from the adult language. To recapitulate, all the conjugational processes discussed in the preceding sections – present, past, and future tense formation, as well as the construction of the imperative and momentary verbs – are more diagrammatic in child language than in adult language. The more irregular, i.e. marked, the process in adult language, the more discernible its diagrammaticity in child language. For example, the formation of the present tense and the imperative, which are highly irregular, entails a lot of overgeneralisations and simplifications in child language with a view to making them more regular, i.e. diagrammatic. On the other hand, the construction of the past tense is quite a regular morphological process and hence it leaves hardly any scope for implementing simplification strategies on the children’s part. Consequently, the degree of its diagrammaticity in child language is only slightly higher than in adult language. The formation of the future tense can be placed somewhere in the middle of the scale, i.e. the degree of its diagrammaticity in child language is neither very low nor very high when compared with adult language. The 150
formation of momentary verbs is an example of yet another issue: its diagrammaticity consists in employing two infixes -nę- and -ną- to mark overtly the momentary aspect of the verb. Having examined various degrees of diagrammaticity in the Polish conjugational system, I will now pass on to a discussion of the phenomenon of diagrammaticity in the Polish declensional system. Diagrammatic nouns The Polish declensional system is extremely sophisticated and irregular (Bąk (1989), Grzegorczykowa (1998)). Its declensional patterns are distinguished on the basis of the noun’s gender. Thus, nouns belong either to the feminine, masculine or neuter declension. The masculine and neuter paradigms are fairly similar when compared with the feminine paradigm, as many neuter and masculine nouns share the same inflectional endings: SG Genitive: Instrumental: Locative:
neuter dzieck-a dziecki-em dzieck-u
– – –
masculine nauczyciel-a nauczyciel-em nauczyciel-u
Another common feature of the masculine and neuter pattern is the great diversity of case endings dependent on the phonological properties of the stem91 and some other factors, e.g. whether the noun is animate or inanimate, virile or nonvirile (Nagórko (1998), Grzegorczykowa (1998)).92 Moreover, within one particular declension or across declensions the so-called inflectional syncretism (Grzegorczykowa (1998)) can be observed which is grammatical ambiguity, i.e. the fact that different cases share the same endings. For example, the inflectional morpheme -a may stand for the inflectional endings of: 1. singular nominative feminine nouns: kobiet-a ‘woman’; 2. singular genitive masculine nouns: pilot-a ‘pilot’; 3. singular accusative masculine nouns: pilot-a ‘pilot’; 4. plural nominative neuter nouns: drzew-a ‘trees’ 91
For example, according to Bąk (1989: 282) the masculine declension comprises nouns whose stems end in: 1. a palatalized consonant: paź ‘page’, 2. a hard consonant: len ‘flax’, 3. a morphological zero: student-ø ‘student’, 4. the vowel -a, characteristic of the feminine declension: kierowca ‘driver’. 92 A detailed discussion of the connection between the type of noun and the inflectional paradigm it belongs to can be found in Bąk (1989: 282–306).
151
The phenomenon of inflectional syncretism goes against the parameters of Natural Morphology, as it violates one of its basic principles, transparency of encoding, which is not preserved in the Polish inflectional system in which one inflectional morph is used to perform a wide variety of functions. Another feature of Polish declensional paradigms is that they are not uniform. Even within one declension, a few inflections mark the same case. Thus, for example, the genitive ending -u is typical for virile inanimate nouns, as in płot-u ‘fence’, while -a is common for virile animate nouns, e.g. kot-a ‘cat’. However, these two groups of nouns exhibit numerous exceptions. Note, e.g. bawoł-u (gen.) ‘buffalo’, or woł-u (gen.) ‘ox’, where animate virile nouns contain the inflectional morpheme that is typical for inanimate virile nouns, or trójkąt-a (gen.) ‘triangle’, where the opposite is the case.93 On the whole, the above mentioned characteristics of Polish declensional paradigms constitute a great impediment to acquiring the system of noun inflectional endings. Hence there is a multitude of instances of confusing inflectional endings in my data, e.g.: child language M (2;2) z tat-em bez koł-ów Pani ma dzidziuś-ø
adult language ‘z tat-ą’, ‘bez kół-ø’, ‘Pani ma dzidziusi-a’
‘with daddy’ ‘without wheels’ ‘The lady has got a baby’
K (2;2–2;4) Nie ma ślad-a do cukr-a Chcę mały kotek-ø
‘Nie ma ślad-u’ ‘do cukr-u’ ‘Chcę małego kotk-a’
‘There is no trace’ ‘into the sugar’ ‘I want a little kitty’
The nouns kotek-ø and dzidziuś-ø here are intended to be in the accusative case and are used instead of the correct adult forms, kotk-a and dzidziusi-a, as a result of confusing the case endings of animate and inanimate nouns. Masculine inanimate nouns have the zero ending in the accusative case, e.g. samolot-ø ‘plane’. Thus, the zero ending has also been adopted here to mark the accusative of animate nouns. The genitive forms ślad-a and cukr-a represent the opposite case, as they have been inflected in accordance with the rules for the masculine, animate gender, whereas the masculine inanimate genitive case should be marked with 93
152
For more discussion concerning the genitive of masculine nouns see e.g. Westfal (1956).
the -u ending: ślad-u and cukr-u. In fact, my data shows that children display a tendency to overextend the -a ending.94 In my opinion, there are two reasons for this phenomenon. First of all, the high frequency of the multifunctional suffix -a95 means that children assume its applicability is much wider than it is. Secondly, the semantic concept of Possessor, which is marked by the use of the genitive case, is acquired by children quite early, much earlier than the semantic concept of Patient, expressed by the accusative case. Since the vast majority of possessors are animate nouns which take the -a ending in the genitive, children overextend it to inanimate nouns. This is because, as my data shows, the form that children learn first serves as the basis for the acquisition of other forms. The instrumental form z tat-em represents an example of the overgeneralisation of the masculine instrumental case of both animate and inanimate case ending characteristic of the vast majority of masculine nouns: z pilot-em ‘with the pilot’, pod samochod-em ‘under the car’, etc. However, the noun tata belongs to a so called mixed declension (Bąk (1989: 291)), which includes nouns of atypical morphological structures, e.g. masculine nouns characterised by the feminine nominative case ending: mężczyzn-a ‘man’, atlet-a ‘athlete’, etc. These nouns take the -ą ending in the instrumental case; hence the correct adult form z tatą. Lastly, the form bez koł-ów represents an example of so-called inflectional imperialism,96 which consists in overextending the plural masculine genitive ending -ów to feminine and neuter nouns. This phenomenon can be accounted for on the grounds of Slobin’s operating principle of overt marking (Slobin (1973)). According to this principle, explicit morphological marking is preferred by children to implicit morphological marking. Thus, affixation is given priority over, for example, modification or suppletion. Hence, the form bez koł-ów, formed by means of affixation, is chosen over bez kół constructed through internal modification. Moreover, applying the -ów ending allows the children to get round the process of alternation, which makes their language more diagrammatic: bez koł-ów versus bez kół, while the nominative plural is koł-a. Alternation constitutes a major obstacle to acquiring the system of noun inflections, in the same way as it considerably slows down mastery of verb inflections. However, Polish nouns vary a lot with regard to the amount of alternation that they exhibit. Nouns from the masculine and neuter declensions 94
Smoczyńska’s research (1985) led her to the same conclusion. However, she is unable to explain this phenomenon. 95 In English the multifunctional suffix -s is also the most often overextended, as has been discussed in chapter II. 96 This term can be found in, for example, Smoczyńska (1985: 627).
153
are characterised by numerous cases of alternation, whereas the nouns belonging to the feminine declension are much more regular in this respect (Grzegorczykowa (1998)). The reason for the greater regularity of the feminine pattern is, first of all, that feminine case endings are far less diverse than masculine and neuter ones (Grzegorczykowa (1998)). Secondly, children first acquire the most prototypical declensional class of feminine lexemes with the -a nominative ending, as the vast majority of feminine nouns that children learn in the second or third year of their lives end with the inflectional morpheme -a in the nominative case, e.g. mam-a ‘mummy’, babci-a ‘granma’, lal-a ‘doll’, piłk-a ‘ball’, zupk-a ‘soup’, czekoladk-a ‘chocolate’, etc. On the basis of my data it can be stated that feminine nouns with other than -a nominative ending are virtually non-existent in children’s speech. However, what is most important is the fact that the declensional paradigm of singular feminine nouns with the nominative -a ending is characterised by only a small degree of alternation.97 Therefore, children overregularise mainly the masculine and neuter patterns both when it comes to overregularising or mixing the inflectional endings and ignoring the necessity of employing alternation. Examples of overregularising the feminine declension have not been observed in my data. As a result, the data on diagrammatic noun forms that I am going to present in the following sections is limited to nouns belonging either to the masculine or neuter declension. Moreover, like in the case of the diagrammatic verbs, the form that children resort to is the most prototypical, i.e. basic one. In this case it is the nominative stem. The reason for this is (as has already been remarked) is that child language is prototype dependent, i.e. it gives priority to basic forms and structures over more sophisticated ones; hence the reliance on the prototypical nominative form. Inflecting nouns using the nominative stem In what follows I will deal with the largest language sample of diagrammatic nouns in my data, i.e. various case forms derived from the nominative stem: 97 The only type of alternation than can be encountered in this pattern occurs in nouns whose stem ends in a non-palatal consonant, e.g. ryb-a ‘fish’ and łąk-a ‘meadow’ (Grzegorczykowa (1998: 275)): Dative: rybi-e, łąc-e Locative: rybi-e, łąc-e The Dative and Locative are the only cases characterised by alternation consisting of consonant palatalization, e.g. b: b’ in ryba-rybie or k: c in łąka-łące.
154
children’s nominative diagrammatic form language K (1;9–2;3); Z (1;8–2;0)) Cukier-u cukier Pantofel-a pantofel Z ludzia-mi ludzie O tydźni-u tydzień Kapcer-a98 Kacper W gniaździ-e gniazdo W laci-e Lato Bóbr-a bóbr Ksiądz-a ksiądz Mąż-a mąż Lew-a Lew Na łokieci-u łokieć Nie umyłaś ręc- ø ręc-e Do przód-a
przód- ø
adult conventional language
‘cukr-u’, i.e. ‘sugar’ (gen.) ‘pantofl-a’, i.e. ‘slipper’ (gen.) ‘z ludź-mi’, i.e. ‘with people’ ‘o tygodni-u’, i.e. ‘about the week’ ‘Kacpr-a’, i.e. ‘Kacper’s’ ‘w gnieździ-e’, i.e. ‘in the nest’ ‘w leci-e’, i.e. ‘in the summer’ ‘ bobr-a’, i.e. ‘beaver’ (dat.) ‘księdz-a’, i.e. ‘priest’ ‘męż-a’, i.e. ‘husband’ ‘lw-a’, i.e. ‘lion’ ‘na łokci-u’, i.e. ‘on the elbow’ ‘Nie umyłaś rąk’, i.e. ‘You have not washed your hands’ ‘do przodu’, i.e. ‘forwards’
98
The form Kapcera, instead of Kacpera is an example of a metathesis. According to Polański (1971), metathesis can be defined as the transposition of sounds within one word. The reason for the occurrence of this phenomenon is that it aims at the facilitation of pronunciation. My data both from Polish and English abounds in numerous instances of metathesis: K (2;6-5;0) petska ‘pestka’ ‘stone’ palpać ‘paplać’ ‘blab’ gmła ‘mgła’ ‘fog’ czworaj ‘wczoraj’ ‘yesterday’ iadro ‘radio’ ‘radio’ u kelarza ‘ u lekarza’ ‘at the doctor’s’ atsygmatyzm ‘astygmatyzm’ ‘astigmatism’ walchuje się ‘wachluje się’ ‘(she is) fanning herself’ ziferek ‘zefirek’ ‘little zephyr’ prokudują ‘produkują’ ‘(they) produce’ pegard ‘gepard’ ‘cheetah’ E (1;8-2;0) It tasty’s ‘It’s tasty’ snall mose ‘small nose’ cumb cratcher ‘crumb catcher’ [ailenz] ‘lions’
155
As can easily be observed, the alternation that is relied on in the adult language but ignored in the child language is either vowel alternation, or both vowel and consonant alternation. Since, according to Smoczyńska (1985), children are capable of performing one phonological operation at a time, they apply just one kind of alternation, e.g. laci-e instead of leci-e from lato, i.e. ‘summer’, or gniaździe instead of gnieździe from gniazdo, i.e. ‘nest’. These forms are not completely diagrammatic because of the consonant alternation but they are more diagrammatic than their adult equivalents because of the lack of vowel alternation. The consonant alternation is preserved here probably because it is fairly regular, as it consists in the palatalization process present in so many other inflectional paradigms. The most diagrammatic noun forms from my corpus of the children’s language are those which in adult Polish merely require vowel alternation, which is ignored by children, e.g. bóbr-a instead of bobr-a from bóbr, as they are formed by the most diagrammatic morphological process, i.e. affixation without even the slightest degree of internal modification. Other examples of this type are mąż-a, ksiądz-a, etc. Another example of vowel alternation that children disregard is e: ø, as in pantofel-a instead of pantofla from pantofel. Thus, in this way they obtain perfectly morphologically transparent, i.e. diagrammatic forms. Other examples of this type are: lew-a instead of lw-a from lew, or Kapcer-a instead of Kacpr-a from Kacper. All in all, the significant richness and complexity of the Polish declensional system, abounding in various types of alternation and declension types, creates favourable opportunities for its simplification and overregularisation. The masculine and the neuter pattern are highly irregular, whereas the feminine pattern is relatively unmarked. As a result, child language is full of diagrammatic masculine and neuter noun forms which are non-existent in the adult language. Having discussed the issue of the formation of diagrammatic nouns using the nominative stem, I will now focus on another iconic process involving nouns, namely pluralising nouns using the nominative stem. Pluralising nouns using the nominative stem As the process of pluralising nouns is more regular than inflecting them, there is far less scope for creating diagrammatic plural nouns than for creating diagrammatic nouns from the nominative stem. Hence the number of diagrammatic plural nouns is smaller than the number of diagrammatic nouns created from the nominative stem. Below are some diagrammatic plural nouns constructed by children. 156
K (2;3–3;5) children’s diagrammatic language uch-y ręk-i ząb-y człowiek-i sen-y rok-i pies-y ok-a
nominative form ucho ręka ząb człowiek sen rok pies oko
adult conventional language ‘uszy’, i.e. ‘ears’ ‘ręce’, i.e. ‘hands’ ‘zęby’, i.e. ‘teeth’ ‘ludzie’, i.e. ‘people’ ‘sny’, i.e. ‘dreams’ ‘lata’, i.e. ‘years’ ‘psy’, i.e. ‘dogs’ ‘oczy’, i.e. ‘eyes’
In order to explain the nature of diagrammaticity in the process of plural noun formation, let me characterise briefly the formation of the plural in adult Polish. Thus, plural nouns are formed by: a) adding the plural nominative ending to nouns whose stems end in ø, e.g. noc ‘night’– noce ‘nights’, dom ‘house’ – domy ‘houses’, etc.; b) replacing the singular nominative case ending with the plural nominative case ending: kobieta ‘woman’ – kobiety ‘women’ woźnica ‘coachman’ – woźnice ‘coachmen’, etc.; c) replacing the singular nominative case ending by the plural nominative case ending accompanied by consonant alternation: oko ‘eye’ – oczy ‘eyes’, ucho ‘ear’ – uszy ‘ears’, etc.; d) adding the plural nominative case ending to nouns which end in ø accompanied by vowel alternation, most often: e: ø, or ą: ę, e.g. sen ‘dream’ – sny ‘dreams’, ząb ‘tooth’ – zęby ‘teeth.; e) suppletion, e.g. człowiek ‘man’ – ludzie ‘people’, rok ‘year’ – lata ‘years’.99 The first way of forming the plural, i.e. (a), is totally diagrammatic as it entails merely affixation. The second method of pluralising nouns, (b), is less diagrammatic because it involves noun modification although the stem remains unchanged. These forms are both morphologically transparent and semantically compositional and their formation does not cause any difficulties for children. Their formation therefore remains beyond the scope of my interest. As far as consonant alternation is concerned – e.g. k: cz in oko ‘eye’ – oczy ‘eyes’ occurring in pattern (c) – children ignore it, resorting simply to replacing the singular nominative case ending with the plural nominative case ending, e.g. ok-a from ok-o ‘eye’ instead of oczy ‘eyes’. In such cases they adopt the adult rule (b) for plural formation. However, lack of consonant alternation entails 99
A detailed list of plural noun inflections can be found in Grzegorczykowa (1998: 273).
157
a change in the inflectional ending denoting plural, which is phonologically conditioned. For example, when the noun stem ends in -k- it cannot take -y as the plural morpheme, because -y does not occur as the plural morpheme after k-; hence the form ok-a.100 Therefore, it can be said that plural nouns formed according to this paradigm are more diagrammatic than adult forms, although they are not totally diagrammatic, as the morpheme boundary is blurred by the vowel alternation,101 e.g. a: i, as in ręk-a ‘hand’ – ręk-i ‘hands’, or o: y, as in uch-o ‘ear’ – uch-y ‘ears’. The forms that are completely diagrammatic in children’s language are those which in adult language require either vowel alternation within the noun stem or suppletion when forming the plural, i.e. they follow the patterns (d) and (e). Children overregularize these plural forms by adhering to the most diagrammatic morphological operation, i.e. affixation without alternation: ząb ‘tooth’ – ząb-y ‘teeth’, pies ‘dog’ – pies-y ‘dogs’, rok ‘year’ – rok-i ‘years’, etc. Consequently, they produce perfectly diagrammatic forms which are not subjected to any kind of internal modification. To sum up, since in adult language the formation of plurals is more regular than the declension process, there are fewer divergences between adult language and child language as far as constructing plural nouns is concerned. This means that some forms which are unmarked, i.e. diagrammatic, in child language coincide with the perfectly diagrammatic forms existing in adult language, e.g. kot-y ‘cats’, dom-y ‘houses’, etc. However, such unmarked forms are in a minority, as the majority of plural nouns are irregular and these forms are overregularised, i.e. made diagrammatic, in child language, e.g. rok-i instead of lata ‘years’, ząb-y instead of zęby, ‘teeth’ etc. The last example of diagrammaticity in the Polish language that I will discuss is the construction of diagrammatic comparative adjective forms. Positive degree oriented comparison of adjectives Diagrammatic comparative adjectives constitute a marginal phenomenon of diagrammaticity in child language. The main reason is that, as my data shows, the comparison of adjectives is a morphological process that is acquired relatively late by Polish children. The first comparative adjectives have been recorded in my data in the speech of children at the age of 2;10. As can be remarked, they are extremely rare: 100
-a is the most common marker of plural for the neuter declension (Grzegorczykowa (1998: 273). 101 In Polish the morpheme boundary is often blurred, as the formation of various noun forms involves affixation followed by internal modification due to the fact that a bare stem is an abstraction, which means not adding but exchanging inflectional endings.
158
children’s diagrammatic language dobr-ejsza lek-sza
positive degree dobra lekka
adult conventional language ‘lepsza’, i.e. ‘better’ ‘lżejsza’, i.e. ‘lighter’
Another reason for the low frequency of diagrammatic adjectives is the fact that in Polish the comparison of adjectives is a fairly regular process. Alternation is therefore not that widespread or extensive as it is in the case of the verb conjugations and noun declensions (Bąk (1989: 242)). As far as my language corpus is concerned, the form dobr-ejsza has been created in accordance with the rules of the regular comparison of adjectives which in the comparative degree take the ending -szy, as in nowy ‘new’ – now-szy ‘newer’ or, -ejszy if the adjective ends in -ny, as in ciemny ‘dark’ – ciemn-iejszy ‘darker’. The choice of the morphological ending -ejszy instead of -szy has been determined by the morphological properties of the stem, i.e. the occurrence of the consonantal cluster -br-, excluding the addition of a morpheme that also begins with the consonant. In adult Polish, however, the adjective dobry ‘good’ (as well as zły ‘bad’, mały ‘small’ and wielki ‘large’) is compared irregularly, i.e. the comparative and the superlative degree are formed suppletively; hence the forms lepszy and najlepszy. As far as the comparative form of the adjective lek-ka ‘light’ (fem.)102 is concerned, it has been formed by adding the correct inflectional morpheme -sza and ignoring the e: ż and k: ej alternation; hence the form leksza instead of lżejsza ‘lighter’. On the whole, the richness and complexity of the Polish morphological system entails its immense irregularity, which, in turn creates ample scope for overregularising many morphological forms in child language, i.e. making them more diagrammatic than those functioning in adult language. The phenomenon of diagrammaticity is especially easily discernible in noun and verb inflection due to the high frequency of nouns and verbs in the children’s speech as well as the significant degree of sophistication of the Polish declensional and conjugational systems. When compared with Polish, the English inflectional system is quite uncomplicated and therefore the phenomenon of overregularising inflectional endings is virtually non-existent. The issue of inflectional diagrammaticity in English will be dealt with in the next section. 4.2.4.2. Inflectional diagrammaticity in English
As has already been stated, a characteristic feature of the English language is its impoverished morphological system. Bound grammatical morphemes are very 102
Adjectival suffixes -ki (masculine) and -ka (feminine) are deleted in the comparative form.
159
few in number and the vast majority of grammatical forms are created regularly, with hardly any alternation. Thus, in contrast with the Polish language the scale of diagrammaticity observed in child language is, in the majority of cases, the same as that in the adult language. This is so because adding -s, -ed, or -ing morphemes to the noun or verb stems does not involve any substantial changes in their pronunciation. The grammatical mistakes made by English-speaking children consist mainly in omitting obligatory grammatical functors,103 not in employing the wrong inflectional patterns or ignoring alternation, which is the case in Polish. Nonetheless, there are two grammatical processes which can be irregular, and which, as a result, are overregularised by English-speaking children. These are the formation of the irregular past and the plural number of nouns which are more diagrammatic in the speech of children than in adult language. Diagrammatic past tense verb forms In this section I will present the way in which children overregularize irregular past verb forms. The majority of first language acquisition researchers (Brown (1973), Cazden (1968), Ervin (1964), Kuczaj (1977), Slobin (1971)) have observed that children overregularise past verb forms by adding the -ed ending to the infinitive stem. However, my data, as well as some child language corpora from the CHILDES project104 clearly show that there are three strategies that children resort to with the aim of making past verb forms more regular, i.e. more diagrammatic: 1. attaching the -ed ending to the infinitive stem children’s diagrammatic language E (2;0) buy-ed catch-ed draw-ed
adult conventional language ‘bought’ ‘caught’ ‘drew’
L (1;5–2;7)105 103
This issue has been dealt with in chapter II devoted to the discussion of the telegraphic speech. CHILDES stands for Child Language Data Exchange System and it is an international project run by Brian MacWhiney aimed at collecting transcripts of children’s language. Speech samples come from children speaking many various languages and have been transcribed in the CHAT format which provides information on context as well as semantic, syntactic and morphological analyses (http://childes.psy.cmu.edu/). 105 L and Jn stand for the names of two girls: Lucy and Jane whose speech has been transcribed as part of the CHILDES project. Both speech samples come from Alan Cruttenden’s file from the Department of Linguistics at the University of Manchester who collected the data from dizygotic twins at monthly intervals (http://childes.psy.cmu.edu/manulas/x45225). 104
160
drink-ed know-ed come-d wake-d break-ed
‘drank’ ‘knew’ ‘came’ ‘woke’ ‘broke’
Jn (1;5–2;7) blow-ed take-d tear-ed
‘blew’ ‘took’ ‘tore’
2. attaching the -ed ending to the stem of the past form of the irregular verb stem children’s diagrammatic language L (1;5–2;7) lost-ed thought-ed broke-d
adult conventional language ‘lost’ ‘thought’ ‘broke’
3. using the auxiliary did as the marker of the past form with either the infinitive or the past verb form. children’s diagrammatic language
adult conventional language
E (2;1–2;2) She did go
‘She went’
Jn (1;5–2;7) I did see I did forget I did not be scared
‘I saw’ ‘I forgot’ ‘I was not scared’
The above three ways of forming iconic irregular past verb forms have been ordered on the basis of their frequency of occurrence in the data. Thus, the most common pattern consists in the addition of the -ed ending to the infinitive stem. This pattern is diagrammatic because the addition of the -ed ending equals the addition of meaning. In contrast, in the adult language the diagrammaticity of irregular verbs is quite low, as they are formed either by modification, e.g. rise-rose, metaphoricity, e.g. put – put, or suppletion, e.g. go 161
– went. The children’s pattern is also in accordance with one of the principles of Natural Morphology concerning the morphological base, which states that the basic forms of words constitute the most natural morphological bases. Another way of forming the past tense, i.e. adding the -ed morpheme to the irregular past form is an example of overmarking. It proves that irregular past verb forms are not morphologically transparent for children; therefore they need the overt marking of the past tense. Another reason for attaching the -ed ending to the past form of the verb may be the fact that children regard the past form as basic since they learned it first. Lastly, the third way of forming the past tense, by means of the auxiliary did, constitutes an example of morphosemantically transparent analytical encoding, as the addition of the past auxiliary equals the addition of the past meaning. On the whole, although the phenomenon of the diagrammatic past is not very widespread in child language due to the limited number of irregular verbs, it constitutes a typical example of inflectional diagrammaticity in English. Iconic plural nouns On the basis of my language corpus it can be said that this phenomenon is marginal in the speech of English-speaking children for two reasons. First of all, irregular plural nouns are very few in number. Secondly, they do not denote basic semantic domains for children in the way that some irregular verbs, such as drank, ate, went, etc. do. Many of the irregular verbs that are overregularised by children are very frequent in the language input as they name basic activities that children or their caretakers engage in. This is not the case with irregular plural nouns, which do not denote basic objects from the children’s point of view and are not commonly used either by children or their caretakers. Because of this, children are virtually deprived of the possibility of overregularising them. Therefore, in my data I have found just three examples of diagrammatic plural nouns. As far as the analysis of the selected CHILDES corpora106 is concerned, I have not come across any diagrammatic plural nouns. Thus, the examples presented below come exclusively from my data: children’s diagrammatic language E (1;8–2;2) mouse-s feet-s fish-es 106
adult conventional language ‘mice’ ‘feet’ ‘fish’
For the purpose of the analysis of plural formation I have investigated the following files: Wooten, Wells, Cruttenden, Fletcher (http://childes.psy.cmu.edu/).
162
All the above plural nouns have been formed by children in the course of the most diagrammatic morphological process, i.e. agglutinative suffixation. As far as the adult forms are concerned, they represent the third and the fourth degree of diagrammaticity, i.e. modification (mice, feet) and metaphoricity (fish) respectively. To repeat what has already been stated, the occurrence of diagrammatic plural nouns is extremely rare in children’s speech because of the very low frequency of irregular plural nouns in the language input. Discussed within the framework of Natural Morphology, diagrammaticity is not the only iconic phenomenon noticeable in the process of plural noun formation. On the basis of my language corpus it can be said that another iconic process, namely metaphoricity, is equally common in the construction of the plural. Here are some plural nouns from my data which are examples of metaphors: plural number in child language E (1;4–2;9) shoe paw hat
plural number in adult language ‘shoes’ ‘paws’ ‘hats’
According to the theory of Natural Morphology, a metaphor is an iconic sign characterised by similarity to its object. This means that a morphological metaphor shows a particular signatum mapped onto a semantically different though morphologically related signans. Taking the above plural forms into consideration, the signatum, i.e. the concept of plurality, is mapped onto a signans that remains morphologically unchanged, although it is semantically different because it denotes the plural. Thus, for example, the noun shoe expressing plurality is a morphological metaphor, as its meaning has altered, although the form remained unchanged. Conclusion As far as inflectional iconicity both in Polish and in English is concerned, its occurrence constitutes conclusive evidence for the urge to implement regularity in language. The more irregular the nature of the language the more intensely this regularity is sought by children. Therefore, inflectional iconicity is much more discernible in Polish, whose morphological structures are characterised by a high degree of irregularity and markedness. Consequently, English, virtually unmarked, as far as its inflectional system is concerned, exhibits marginal inflectional iconicity, resulting from oversimplification of grammatical forms. 163
3.2.5. Derivational iconicity So far I have been discussing the issue of iconicity in inflectional morphology. However, iconicity is also encountered in the process of word formation, both in English and in Polish. I am going to investigate the issue of iconicity in derivational morphology on the basis of deriving innovative nouns and verbs. The reason I choose to tackle nouns in my analysis is that in any world language they constitute the largest percentage of all coinages created by children because labels for objects are the most frequent items in their speech (Chmura-Klekotowa (1971), Clark (1987)). Verbs are the second largest group of morphological neologisms both in English and Polish.107 Since the phenomenon of coining innovative nouns and verbs is tantamount to naming objects and actions, I am also going to analyse the cognitive factors responsible for this process. According to G. Radden and K. Panther (2004), the process of naming is also guided by language-independent factors, such as salience, economy and metonymy. Thus, only certain components of the complex concept are selected and named, as will be substantiated in the following section. It will also be shown that innovative nouns and verbs coined both by Polish and English-speaking children are iconic, as they exhibit a resemblance to the objects or actions from which they have been derived, whereas their adult counterparts are most often arbitrary ones. For example, an innovative noun from my data stukacze has been derived from the verb stukać ‘clatter’ to denote high-heeled shoes, making a clattering noise. On the basis of this particular derivative it is evident the process of its coining can be characterised by: 1. salience – the most perceptually salient feature has been noticed by the child; 2. economy – the choice of lexical means has been reduced here to the minimum, whereas in adult language the meaning of this single word would have to be expressed by means of a complex phrase, e.g. high-heeled shoes that make a clattering noise; 3. metonymy – the coinage stukacze is a reflection just of one of the features of high-heeled shoes, namely the fact that they make a clattering noise; this derivative has been motivated by a PART FOR WHOLE metonymy; 4. iconicity – the meaning of the noun can be related to the meaning of the verb it has been derived from; 5. diagrammaticity – the addition of the nominal suffix -acz equals the addition of meaning: ‘the object V-acz is capable of V-ing’, where V stands for a verb.
107
A detailed discussion of morphological neologisms in children’s speech from Polish and English data may be found in Chmura-Klekotowa (1971) and Clark (1987).
164
3.2.5.1. Iconic innovative nouns in English 108
In the early period of the linguistic development of English-speaking children, compounding is the most productive morphological operation for coining new nouns (Szymanek (1998), Clark (1987)). Compounds produced by children can be divided into root compounds, formed from two or more nouns, as in house-key, and synthetic compounds, formed from one or two nouns combined with a verb, e.g. push-chair (Clark (1993)). All the innovative nouns from my data except for the last two are examples of root compounds: iconic compound J (2;6–3;10) bear-hat clown-boy bear-son
adult counterpart ‘the bear that looks like a hat’ ‘the boy who is a clown’ ‘the son who is a bear’
E (1;7–2;9) baby-elephant baby-bottle farmers-market daddy-seed crumb-catcher mama-bunny Ringo grocery boy-lion bobo-tea water-boat pee-pee cup hangie
‘the elephant that is a baby’ ‘the bottle meant for babies’ ‘the market meant for farmers’ ‘the seed that is a daddy’ ‘the object capable of catching crumbs’ ‘the bunny that is a mama’ ‘the grocery that belongs to Gringo’ ‘the lion that is a boy’ ‘the tea for a baby’ ‘the boat floating on water’ ‘the cup meant for peeing’ ‘a hanging toy’
All the above compounds are examples of one way in which children code messages they want to express. According to Heine (1997), the content of the message should be presented with clarity and ought to be coded in such a way that the hearer can interpret it with minimal processing effort. In other words, economic motivation is of primary importance. This is the case with the compounds formed by children, e.g. the compound baby-bottle can easily be decoded as ‘the bottle meant for/used by the baby’, baby-elephant can easily be 108
The contents of this subchapter has served as the basis of analysis of innovative nouns created by metonymy that has been published as an article entitled ‘Motivation in language: the case of metonymically motivated innovative nouns in children's speech’ in 2006 in the linguistic journal Odisea: Revista de Estudios Ingleses (pp.103–118).
165
understood as ‘the elephant who is a baby’, etc. What is more, these compounds are relatively short, much shorter than their adult counterparts, e.g. bobo-tea for ‘tea meant for babies’. However, the question arises: what drives children to code their messages in such a way? My attempt to account for the cognitive mechanisms responsible for the process of compound formation will be based on the phenomenon of apperception (Rozwadowski (1903), Tabakowska (2004)). Cognitive psychology defines apperception as the process in the course of which the attention of the individual is focused on a single point. All other potentially perceptible elements lingering in the field of our consciousness are referred to as perceptions. (Hutchinson Encyclopaedia (2000)). Although the concept of apperception has won widespread popularity only recently, with the advent of cognitivism, it was discovered for linguistics as early as at the beginning of the 20th century by Jan Rozwadowski, one of the most eminent Polish linguists, whose ideas have given rise to the modern cognitive theories of language. Rozwadowski’s theory was strongly influenced by Wundt (1900), according to whom the process of perception consists in the ability to distinguish in every word, phrase or sentence the so-called identifier and diversifier, where the identifier is the identifying element and the diversifier is the distinguishing element. Thus, in the phrase blue dress, the identifier is dress, as it classifies the object within the class DRESS, whereas the diversifier is blue because it distinguishes this particular object form other kinds of dresses. According to Rozwadowski, the perception of things and phenomena defined as apperception consists in the perception of the changing reality, accompanied by the relation of new facts to the facts that are already known from previous knowledge and experience. Several modern linguists have made use of the definition of apperception put forward by Rozwadowski. For example, according to Klimkowski (2005), apperception is the manifestation of iconicity and therefore it is possible to talk about so-called apperceptive iconicity because the process of the formation of words and phrases mirrors the cognitive strategies of encoding their meaning. Klimkowski has made use of Rozwadowski’s theory to analyse the process of compound formation in English. He notes that according to Rozwadowski the process of phrase formation comprises two cognitive strategies: 1. establishing similarity links between already existing utterances (and their meanings) 2. establishing points of difference between already existing utterances and new ones. Analysing the root compounds from my data presented above, I am going to rely on the above mentioned assumptions of the theory of apperception. 166
Thus, it is undeniable that the root compounds from my data are all constructed in accordance with the principles of apperceptive iconicity. This means that their second element is always the identifier, as it establishes similarity links between the already existing utterances, i.e. it places objects within a certain class: daddy-seed crumb-catcher baby-elephant
this object belongs to the class SEED; this object belongs to the class CATCHER; this object belongs to the class ELEPHANT,
whereas their first element is the diversifier, as it establishes points of difference between already existing utterances and new ones, i.e. it distinguishes these particular objects from other kinds of objects: daddy-seed not e.g. mummy-seed, i.e. this object is daddy not, e.g. mummy; crumb-catcher not e.g. fly-catcher, i.e. the function of this object is to catch crumbs not, e.g. flies; baby-elephant not e.g. boy-elephant, i.e. this object is a baby not, e.g. a boy. As far as the other type of compounds is concerned, in my data there is just one example of a synthetic compound, i.e. pee-pee-cup where cup is the identifier and pee-pee is the diversifier.109 All the compounds, both root and synthetic, coined by children are iconic, as their structure reflects cognitive mechanisms adopted by children during their formation. In other words, the relationship between the modifying and classifying element is manifested in the form of the compound in which its second part, the identifier, represents the concept known from the child’s previous knowledge and experience, whereas its initial part, i.e. diversifier is always related to the child’s new experience or changing aspect(s) of reality. In my data the only innovative noun which is not an example of a compound is the deverbal nominalisation hangie, meant to denote a hanging toy. Affixation is not a productive process of coining innovative nouns in English (Clark (1987)); hence, innovative nominalisations in English are extremely rare.110 In Polish, unlike in English, deverbal nominalisation constitutes a fairly productive process of innovative nouns formation (Chmura-Klekotowa, (1967)). This morphological operation will be discussed in the next section. 109
In Clark’s data synthetic compounds constitute just 12 per cent of all the compounds. According to Clark (1987), derivation is responsible for the formation of 20 per cent of innovative nouns among children between 2;0 and 3;11. 110
167
3.2.5.2. Iconic innovative nouns in Polish 111
As far as innovative nouns in my data are concerned, the vast majority of them, about 90 per cent, are deverbal nominalisations. Chmura-Klekotowa (1967) does not specify the percentage of deverbal nouns in her data, but concludes that their frequency is relatively high. They are formed to denote agents, instruments, objects and outcomes of activities. Let me now present numerous examples of innovative nouns from my data: a. agents: iconic apperceptive verbal base symbolic adult term nominalisation obcina- czka obcinać, i.e. ‘to cut’ ‘fryzjerka’, i.e. ‘hairdresser’ zapomi-nek zapominać, i.e. ‘to forget’ ‘ktoś kto zapomina’, i.e. ‘somebody who is forgetful’ aresztow-nik aresztować, i.e. ‘to arrest’ ‘aresztujący policjant’, i.e. ‘policeman who is arresting’ remontow-nik remontować, i.e. ‘to ‘pracownik brygady renovate’ remontowej’, i.e. ‘worker of renovation team’ równi-arz równać, i.e. ‘to level’ ‘robotnik równający drogę’, i.e. ‘worker levelling the road’ rozkaź-nik rozkazywać, i.e. ‘to order’ ‘majster’, i.e. ‘foreman’ wykluwa-czek wykluć się, i.e. ‘to hatch’ ‘pisklę wyklute z jajka’, i.e. ‘hatched chick’ b. instruments: iconic apperceptive verbal base nominalization miesza-czka mieszać, i.e. ‘to stir/to mix’ wyciera-czka wycierać, i.e. ‘to dust’ ściera-czek wyłącze-nie zasuwa-nka nawij-ka
symbolic adult term ‘łyżka’, i.e. ‘spoon’
‘miotełka do wycierania kurzu’ ‘feather duster’ ścierać, i.e. ‘to wipe’ ‘mop’, i.e. ‘mop’ wyłączyć i.e. ‘to turn off’ ‘wyłącznik’, i.e. ‘switch’ zasuwać, i.e. ‘to do up’ ‘zamek błyskawiczny’, i.e. ‘zip’ nawijać, i.e. ‘to roll up’ ‘korba przy studni’, i.e. ‘well crank’
111 Innovative Polish nouns in child language have also been discussed in the paper ‘Derivational neologisms in children's speech from Polish and English data’ in 2002 in SAR 1 (pp. 52–64).
168
klicz-ka przekłuj-nik naucz-nik podgląda-czek
klikać, i.e. ‘to click’
‘myszka od komputera’, i.e. ‘computer mouse’ przekłuwać, i.e. ‘to prick’ ‘igła do przekłuwania uszu’, i.e. ‘needle for pricking ears’ nauczyć, i.e. ‘to teach’ ‘zestaw podręczników’, i.e. ‘a set of textbooks’ podglądać, i.e. ‘to peep’ ‘judasz’, i.e. ‘peephole’
c. objects: iconic apperceptive verbal base nominalisation trzyma-nka trzymać, i.e. ‘to hold’ turla-czka turlać się, i.e. ‘to roll’ wysuwa-nka pie-czeń złapa-nie myślo-nka stuka-cze
pokazu-nek
symbolic adult term
‘poręcz’, i.e. ‘hand-rail’ ‘(turlająca się) szpula’, i.e. ‘rolling reel’ wysuwać, i.e. ‘to pull out’ ‘wysuwana półka’, i.e. ‘pulled out shelf’ piec, i.e. ‘to burn’ ‘piekąca rana’, i.e. ‘burning wound’ złapać, i.e. ‘to get hold of’ ‘uchwyt meblowy’, i.e. ‘handle’ myśleć, i.e. ‘to think’ ‘mózg’, i.e. ‘brain’ stukać, i.e. ‘to clatter’ ‘buty na obcasach, które stukaja’, i.e.‘high-heeled clattering shoes’ pokazywać, i.e. ‘to show’ ‘rysunek’, i.e. ‘drawing’
d. outcomes of activities: iconic apperceptive verbal base nominalisation zawiąza-nie zawiązać, i.e. ‘to tie’ pis-ki pisać, i.e. ‘to write’
symbolic adult term ‘kokarda’, i.e. ‘bow’ ‘wydruk z komputera’, i.e. ‘computer printout’
According to Chmura-Klekotowa (1967), the source of these neologisms stems from the development of cause and effect thinking in children aged between two and three. This, in turn, entails children’s interest in the etymology of newly acquired words. What is of primary importance for the present considerations, is the fact that to children’s knowledge every name is not only closely related to its designatum but also refers to one of the activities that these objects are somehow associated with. Hence, all the deverbal nominalisations from my data pertain to: 169
1. the activities that the agents (a) and instruments (b) engage in, e.g. rozkaźnik is somebody who gives orders and commands to people working on a building site (adult ‘majster’, i.e. ‘foreman’), miesza-czka is a piece of cutlery used for stirring or mixing (adult ‘łyżka’, i.e. ‘spoon’); 2. the activities that the objects (c) are meant for, e.g. wysuwa-nka is something that is meant to be pulled out (adult wysuwana półka’, i.e. ‘pulled out shelf’); 3. the outcomes of the activities (d) performed on objects, e.g. zawiąza-nie is something that has been tied (adult ‘kokarda’, i.e. ‘bow’). Therefore, it can be said that all the above coinages are iconic as they exhibit a close resemblance to the objects that they denote because they point directly to the functions that, in the children’s opinion, are the most characteristic of them. Thus, stukacze are shoes whose primary function is that of clattering, pokazunek is a drawing whose primary function is that of showing things, wykluwaczek is a chick whose primary function is to hatch, etc. In order to explain the cognitive strategies used by children in the process of forming innovative nouns I will rely once again on the theory of apperception. As already stated, according to Rozwadowski (1904), apperception consists in relating new facts to facts that are already known from previous experience. The nature of this process makes us perceive distinctly only a limited number of elements out of the total amount lingering in the field of our perception. Because of this, the word becomes a reflection of merely one idea out of all the ideas that it represents. For Rozwadowski it is the dominant feature that becomes embodied in a particular word. As far as the very process of naming objects is concerned, Rozwadowski claims that it consists in highlighting a single feature out of the many features that characterise a given object. Rozwadowski states that every word always denotes a complex and sophisticated notion; however, during the process of naming just one element of this notion is perceived as dominant, while all the remaining elements are considered to be far less distinctive.112 Therefore, all the nominalisations from my data are apperceptive as they are formed by highlighting that function of the object or person that the children perceive as dominant in the process of naming particular objects or people, e.g.: turla-czka, i.e. ‘reel’ – the dominant feature of this object is that it is capable of rolling, not that thread can, for example, be wound around it; 112
Needless to say, this highlighting is known in modern terminology as salience (Tabakowska (2004), Slobin (1985)). What is more, Rozwadowski’s ideas based on the concept of duality in language gave rise to modern cognitive theories of language. Both Langacker’s distinction between the profile/figure and base/ground and the cognitivist theory of metaphor are deeply rooted in Rozwadowski’s thinking.
170
obcina-czka, i.e. – the dominant feature of this person is that she cuts ‘hairdresser’ people’s hair, not, for example, that she dyes it; zasuwa-nka, i.e. ‘zip’ – the dominant feature of this object is that it is used for doing up things, not that it is for example made of metal or plastic. Evidently, there are close links between Rozwadowski’s theory and the modern cognitivist theory of metaphor and metonymy. The reason for this is that the theory of apperception is capable of accounting for the nature of metaphorical and metonymical processes and operations by emphasising the fact that in the naming process our attention is concentrated on the most salient feature of a given object which, afterwards, serves as a basis for coining a new word or phrase. Assuming that metonymy expresses a PART FOR WHOLE relationship,113 it may be said that the nouns from my data can also be labelled as metonymical derivatives. They are mentonymical in the sense that the meaning they denote constitutes just a part of the whole range of meanings that a particular noun conveys. Thus, it can be said that the above nouns from my data tend to be motivated by PART FOR WHOLE metonymy. For example, the noun zawiązanie ‘bow’ from the verb wiązać, i.e. ‘tie’ is metonymical in the sense that it expresses just a part of the whole composite concept that the noun bow denotes. This coinage refers to just one of the characteristic features that a bow may have, namely that it can be tied. However, this noun represents the composite concept by which its general characteristics should be understood, such as the material it is made of, its colour, its width, the purpose for which it is going to be used, the way it is going to be fastened, etc. These features have not been included in the coinage, and so it can be referred to as metonymical. Thus, highlighting the dominant functions of a given object or a person entails ignoring those functions that are considered to be less distinctive. This leads to metonymically motivated phenomena in the process of word-formation. The cognitive process of apperception turns out to be responsible for coining innovative nouns both in English and in Polish, although the morphological operations employed in each case are entirely different. Englishspeaking children resort to compounding, whereas Polish children make use of deverbal nominalisation. The types of derivational mechanisms that Polish and English-speaking children resort to are merely reflections of the word-formation strategies relied on in the process of forming nouns in adult language. However different these processes are in the two languages, they have one feature in common – their underlying mechanism is apperceptive iconicity. Apperceptive 113
See Lakoff and Johnson (1980), and Nerlich and Clarke (2001).
171
iconicity is also responsible, at least to some extent, for the process of innovative verb formation, both in Polish and English, which will be discussed in the following section. 3.2.5.3. Iconic innovative verbs in English
As far as English innovative verbs are concerned, they constitute a considerably smaller percentage of all the utterances than innovative nouns. In my corpus their occurrence is marginal, as there are just three examples of innovative verbs derived through zero-derivation: E (1;7–2;9) iconic innovative 1. argument 2. sharp 3. stroll
verb adult term ‘argue’ ‘sharpen’ ‘carry sb in the stroller’
The above examples have been formed by: 1. denominal zero-derivation from the noun argument 2. deadjectival zero-derivation from the adjective sharp 3. back-derivation from the noun stroller According to the theory of Natural Morphology, all types of morphological conversion are metaphors, as a metaphor is an iconic sign characterised by similarity to its object. The first two verbs from my list, sharp and argument, are examples of metaphors. The similarity of these verbs to the nouns they have been derived from consists in the sameness of the form and close semantic relationship between the meaning of the noun base and the verb,114 accompanied by a difference in the assignment to the grammatical category. With regard to the innovative verb stroll, it is an example of a diagram formed through the process of subtraction which consists in the modification of the lexical base by the deletion of phonological material. Thus, all the verbs from my data are iconic signs, which is not the case in adult language. However, my sample of innovative verbs is too small to constitute a reliable source of reference; therefore, in my considerations, I will rely on the language corpus gathered by E. Clark (1998) who has carried out extensive research into lexical innovations in children’s speech. She has discovered that in English the formation of verbs from nouns through conversion is the most productive, as it constitutes 72 per cent of all the coinages. The second most popular method of 114
The nature of the semantic relationship between the noun and the verb derived from it will be explained further in this section.
172
verb formation is, according to her, deadjectival zero-derivation. Now let me present some examples of morphological metaphors from Clark’s data: iconic denominal innovative verb scissor axe wheel marble wrench needle trouser broom match lemonade
adult counterpart ‘open with scissors’ ‘tap with an axe’ ‘make wheels turn’ ‘roll like marbles’ ‘undo with a wrench’ ‘mend with a needle’ ‘put on trousers’ ‘hit with a broom’ ‘light with a match’ ‘pour lemonade’
iconic deadjectival innovative verb tall sore dark pink
adult counterpart ‘get taller’ ‘feel sore’ ‘make dark’ ‘make pink’
The above listed innovative metaphorical verbs exhibit similarity to the words they have been derived from. Not only is their form identical to the nouns or adjectives they have been derived from, but also they show a close resemblance to their meaning. Iconic denominal verbs are derived from the nouns denoting the names of the instruments by means of which the action is performed, e.g. needle, or wrench, or the objects on which the action is performed, e.g. wheel or trousers. Iconic deadjectival verbs are derived either from adjectives by which the subject is characterised, e.g. tall or sore, or from adjectives denoting the features that the subject is capable of bringing about, e.g. dark or pink. Thus, there exists a close resemblance between each derived verb and its base. When talking about coining innovative verbs, the process of reversative verb formation is worth mentioning as well, as it is highly iconic. According to Clark (1993), children opt for two methods of forming verbs for undoing actions. Young children (between two and three) use particles to express reversal; however, as this is not a productive way of coining verbs for undoing actions, older children overgeneralise the negative prefix un-, using it quite indiscriminately. First of all, let me analyse some instances of expressing reversal by means of particles: 173
iconic innovative verb sink up button down stand down tuck out
verbal base sink down button up stand up tuck in
adult counterpart ‘rise to the surface’ ‘undo’ ‘lie down’ ‘untuck’
As can easily be remarked, the underlying principle of this type of reversative verb formation is the principle of contrast. According to ChmuraKlekotowa (1967), young children are characterised by an aptitude for observing opposing features and meanings, which leads to the juxtaposition of prefixes performing a distinctive semantic function. Consequently, the verb sink up has been coined to convey the opposite meaning to the verb sink down; button down has been coined to express the opposite to button up, etc. Thus, replacing one particle with its opposite is tantamount to changing the meaning of the whole prepositional verb into its opposite. Therefore, such a reversative verb is iconic, as its signans shows a relation of analogy with its signatum, i.e. the addition of the opposite signans makes the whole signatum acquire the opposite meaning. Nonetheless, the above case is merely that of contextual iconicity, as these reversative verbs are iconic only when compared with the verbs they have been derived from. If deprived of the context, their iconicity would not be that apparent. For example, the meaning of the verb stand down taken out of verbal and situational context would give rise to a variety of interpretations. However, if assumed that the child has said I am standing up when rising form a horizontal position the utterance I am standing down can be interpreted as a description of the opposite process, i.e. ‘I am lying down’. In other words, when interpreted in the context, the meaning of such a complex verb is semantically compositional, i.e. its meaning is a function of the meaning of its constitutive parts, e.g. button + down = ‘fasten something by pushing its buttons through buttonholes’ + ‘do something opposite to what was done previously, i.e. buttoning up’ = ‘undo’ tuck + out = ‘put a piece of clothing into a particular place or position’ + ‘do something opposite to what was done previously, i.e. tucking in’ = ‘untuck’ Evidently, this way of forming reversative verbs is possible thanks to the fact that children have learnt that the pairs of particles, such as in – out, up – down are used to express contrasting meanings. 174
Another way of forming reversative verbs, which is much more productive than the use of particles, is the addition of the prefix un-. Clark’s data (1993: 227) abounds in examples of this type: iconic innovative verb un-tight un-disappear un-build un-hang un-crumple un-blow un-string un-light
verbal base tight disappear build hang crumple blow string light
adult counterpart ‘loosen’ ‘reappear’ ‘take apart’ ‘take down’ ‘smooth out’ ‘deflate’ ‘take off’ ‘extinguish’
On the basis of the above examples it is evident that reversative verbs formed by the addition of the negative prefix un- are perfectly diagrammatic because they represent the first degree of diagrammaticity, i.e. agglutinative affixation. They are both morphotactically transparent and semantically motivated, as the addition of the prefix un- equals the acquisition of reversative meaning. Consider the following example: un + hang = ‘put something in a hanging position’ + ‘do something opposite to what was done previously, i.e. hanging’ = ‘take down’ As is easily remarked, reversative verbs functioning in child language are iconic, whereas their adult counterparts are formed suppletively, i.e. they are arbitrary. Their symbolicity stems from the fact that there is nothing in the derived verb that would resemble in any way its base, as in, for example, the pair of opposites: hang – take down. On the whole, in English two kinds of innovative verbs coined by children can be distinguished. First of all, there is either deadjectival or denominal verbalisation. Secondly, there is the formation of reversative verbs by means of either the addition of particles or the prefix un-. Whatever process is employed by the child, the verb derived in the course of it is much more iconic than its adult counterpart. 3.2.5.4. Iconic innovative verbs in Polish
As far as innovative verbs coined by Polish children are concerned, prefixation plays a crucial role in this process, as in Polish prefixes are of
175
primary importance for verb formation because they express a wide variety of both semantic and aspectual differences.115 On the basis of my data three ways of forming verbal neologisms dependent on prefixation can be differentiated: omitting an initial syllable considered to be the perfective prefix, e.g. grywa for ‘na-grywa się’, i.e. ‘it is being recorded’; confusing prefixes, e.g. za-kasować for ‘s-kasować’, i.e. ‘punch’; and forming reversative verbs by using prefixes conveying the opposite meaning, e.g. wykładać from wkładać, i.e. ‘put in’ for ‘wyciągać’, ‘take out’. The cognitive mechanisms responsible for the formation of these iconic verbs can be accounted for on the grounds of the theory of apperception. Thus, it is the verb stem that children’s attention is focused on, whereas prefixes seem to be treated as ‘less important’, and as a result tend to be either deleted (1), confused (2), or misused (3).116 In other words, the verb stem is considered to be the identifier and the prefix the diversifier. As already stated, the identifier places the word, in this case the verb, within a certain class, whereas the diversifier distinguishes this verb from other verbs belonging to the same class. For example, in one of the coinages produced by Polish children, wy-wiązać for ‘odwiązać’, i.e. ‘undo’, the identifier is wiązać ‘tie’, because it places this verb within the class TIE, whereas wy- is the diversifier because it makes the verb wiązać different from other verbs, such as przy-wiązać ‘tie to’, za-wiązać ‘tie up’, po-wiązać ‘tie together’, prze-wiązać ‘tie with’, etc.117 In what follows I am going to present three groups of innovative verbs in which prefixation plays a crucial role: 1. omitting an initial syllable considered to be the perfective prefix: iconic innovative verb glądać bijać waliła mienimy prawić 115
verbal base, i.e. adult counterpart118 o-glądać, i.e. ‘watch’ u-bijać, i.e. ‘beat’ roz-waliła się, i.e. ‘it crumbled’ za-mienimy się, i.e. ‘we will make an exchange’ na-prawić, i.e. ‘mend’
A discussion of the nature of prefixes in Polish can be found in, for example, Bąk (1989: 259) and Fisiak (1978: 107). 116 These three groups of iconic innovative verbs have been listed below. 117 As I have already remarked, the theory of apperception has given rise to the distinction between the profile and the base, put forward by Langacker. Consequently, the verb stem in children’s iconic innovations is the profile, whereas the prefix is the base (Langacker (1991)). 118 In this particular case the verbal base is the same as the adult counterpart, as Polish children resort here to subtraction, i.e. they delete the initial syllable of verbs functioning in adult language.
176
kręcam lewam stawić prasza przątam
od-kręcam, i.e. ‘I am twisting off’ prze-lewam, i.e. ‘I am pouring’ po-stawić, i.e. ‘put’ prze-prasza, i.e. ‘he is apologising’ s-przątam, i.e. ‘I am tidying’
The source of this kind of lexical innovation goes back to the phenomenon of improper segmentation of input, which is very common in children’s speech.119 Children either regard two words as one word, as in the sentence Dlaczego sójka odwiedziła po-ciotunię? ‘Why did the jay visit her [after-aunt]’;120 or do the opposite, i.e. divide indivisible words into two parts, which is the case above. For example, every verb from this sample is treated not as a whole, but as a prefix and stem. For instance, the verb oglądać is believed by children to consist of the prefix o- and the stem -glądać; the verb ubijać is considered to be composed of the prefix u- and the stem -bijać, etc. The initial syllables of all these verbs are regarded as prefixes on the grounds of one of the principles of Natural Morphology, namely the transparency of encoding. Transparent encoding occurs if a given signans represents one and only one signatum. To children’s knowledge, the initial syllable of the verb is the signans, whose only function is to represent the prefix. As the prefix is the diversifier, it merely modifies the meaning and grammatical form of the verb and is therefore omitted in children’s speech, which is characterised by the tendency to disregard both grammatical and semantic subtleties. Hence, omission of the initial verb syllable can be observed in the speech of the Polish children. Within the framework of Natural Morphology this process is referred to as subtraction, as it involves the deletion of the morphological material and therefore represents the least natural degree of diagrammaticity. The diagrammaticity of these verbs stems from the fact that subtraction of the initial syllable believed to be a prefix equals subtraction of meaning. What I mean by this is that, to children’s knowledge, these verbs become imperfective once their ‘prefixes’ are taken away. 2. confusing prefixes iconic innovative verb u-rwij o-strugaj
adult counterpart ‘wy-rwij’, i.e. ‘tear out’ ‘za-strugaj’, i.e. ‘sharpen’
119
On this issue see Chmura-Klekotowa (1967). The origin of this neologism should be traced back to the following part of a rhyme by Jan Brzechwa: Po ciotuni jeszcze sójka/Odwiedziła w mieścia wujka, ‘After the visit at her aunt the jay called on her uncle’. 120
177
wy-sycha wy-lepić wy-wiązać za-suń za-plamiłam prze-raziłam wy-klej
‘u-sycha’, i.e. ‘it withers’ ‘u-lepić’, i.e. ‘mould’ ‘od-wiązać’, i.e. ‘undo’ ‘przy-suń’, i.e. ‘move (it) closer’ ‘po-plamiłam’, i.e. ‘I have stained (it)’ ‘za-raziłam się’, i.e. ‘I have become infected’ ‘od-klej’, i.e. ‘unstick’
The grounds for this process are both grammatical and semantic, as in Polish verbal prefixes, apart from functioning as markers of perfectivity, frequently also convey a distinct meaning. For example, all the prefixes from the above sample, o-, u-, wy-, za-, prze-, po-, od- and przy- are not only used to denote completed actions, but many of them also express an independent meaning. The meaning of some of these prefixes is as follows: przezawyodpoprzy-
‘across’, ‘through’ ‘completely’, ‘as far as possible’ ‘from the inside’, ‘away from’ ‘away’, ‘away from’ ‘a few times’ ‘to’, ‘in the direction of’
As already stated, the iconicity of the verbs from this set is both grammatically and semantically motivated. It is grammatically motivated in the sense that all the verbs produced by children are meant to express completed actions and for that purpose the children studied used perfective prefixes. What is iconic in this particular aspect of verb formation is the fact that all the verbs are still diagrammatic although incorrect from the adult point of view because incorrect prefixes have been attached to the verb stems. In other words, these verbs are diagrammatic because the addition of the prefix equals the addition of grammatical meaning: wy-klej = ‘complete the activity expressed by the verb’ + ‘use the glue’ o-strugaj = ‘complete the activity expressed by the verb’ + ‘sharpen’, etc. As far as the semantics of these verbs is concerned, the meaning of prefixes attached by children to the verb stems is similar to the meaning of the verbal prefixes used by adults: 1. za-plamiłam (child language) = ‘completely’ + ‘I have stained myself’, i.e. ‘I have stained myself’ completely for po-plamiłam się (adult language) = 178
‘a few times’ + ‘I have stained myself’, i.e. ‘I have stained myself a few times’, or 2. za-suń (child language) = ‘as far as possible’ + ‘move’, i.e. ‘Move (it) as far as possible’ for przy-suń (adult language) = ‘in the direction of’ + ‘move’, i.e. ‘Move (it) in the direction of’ Thus, all these verbs are both morphotactically and morphosemantically compositional because the addition of a prefix brings about a grammatical and semantic change to the structure of the verb, which, as a result, becomes perfective and acquires a different shade of meaning. For example, the addition of the prefix wy- to the verb -klej makes the whole verb perfective and semantically different from the verb klej ‘stick’. As the confusion of these prefixes is semantically and grammatically motivated, the meaning of the whole verb coined by children can not only be conjectured but also is close to that functioning in the adult language. 3. forming reversative verbs by using prefixes conveying the opposite meaning: iconic innovative verbal base verb od-mieszam mieszam, i.e. ‘I will mix’ od-farbowałam farbowałam, i.e. ‘I have dyed’ od-włożysz włożysz, i.e. ‘You will put in’ od-szyło się
szyło się, i.e.‘It has sewn itself’
adult counterpart ‘posegreguję’, i.e. ‘I will sort’ ‘usunęłam farbę’, i.e. ‘I have removed the dye’ ‘wyciągniesz’, i.e. ‘You will take out’ ‘odpruło się’, i.e. ‘It has come off’
The underlying principle of this kind of innovative verb formation in adult language goes against the principle of transparency of encoding, as reversative verbs are formed suppletively. As far as the children’s approach in this respect is concerned, the only way of expressing reversal is to attach the prefix od- to the verb stem. This method of reversative verbs formation stems from the fact that many verbs form their opposites by being prefixed with od-, e.g. dać ‘give’ nieść ‘carry’ jechać ‘go’
– – –
od-dać ‘give back’ od-nieść ‘carry back’ od-jechać ‘go away’
Hence, by analogy with this type of reversative verbs a significant number of verbs expressing reversal are formed by means of the reversative prefix od-. 179
These innovative reversative verbs created by children are diagrammatic because the addition of the prefix od- equals the addition of reversative meaning, e.g. od-włożysz = ‘do something opposite to the activity expressed by the verb’ + ‘put in’ for ‘wyciągniesz’, i.e. ‘you will take out’. Since in adult language reversative verbs are most often formed suppletively, it can be stated that the notion of reversal is much more iconic in child language than in adult language. As far as innovative verbs are concerned, apart from creating prefixation oriented verbal neologisms, Polish children also resort to metaphorical denominal verbalisation: iconic innovative verb grzebieniować się zaguzikować kosmetykować się komputerować wózkować wilczyć rowerować
nominal base
adult counterpart
grzebień, i.e. ‘a comb’
‘czesać się’, i.e. ‘comb one’s hair’ guzik, i.e. ‘button’ ‘zapiąć guziki’, i.e. ‘button up’ kosmetyk, i.e. ‘cosmetic’ ‘robić makijaż’, i.e. ‘make up’ komputer, i.e. ‘computer’ ‘pracować na komputerze’, i.e. ‘use a computer’ wózek, i.e. ‘pram’ ‘pchać wózek’, i.e. ‘push a pram’ wilk, i.e. ‘wolf’ ‘wyć jak wilk’, i.e. ‘howl like a wolf’ rower, i.e. ‘bike’ ‘jechać na rowerze’, i.e. ‘ride a bike’
I would like to advance the thesis that from a semantic point of view these iconic innovative verbs are metaphors because they are characterised by their similarity to the objects they have been derived from. Thus, the verb za-guzikować ‘button up’ is metaphorical because it exhibits resemblance to the object it denotes. It has been derived from the name of the object that the activity is performed on, i.e. guzik, ‘button’, and it names the activity performed on this object. If the verb kosmetykować się ‘make up’ is considered, it has been derived from the name of the object used as an instrument in the process denoted by the verb, i.e. kosmetyk, ‘cosmetic’, and it names the activity performed by means of this object. On the whole, the process of innovative verb formation is much more uniform than the process of innovative noun formation as far as the use of 180
derivational mechanisms both in Polish and English is concerned. In both languages the dominant strategies that children rely on are prefixation and metaphoricisation. Prefixation is not only more diverse in Polish than in English due to the more developed and sophisticated system of verbal prefixes, but it is also relied on to a greater extent. In contrast, English-speaking children recourse most often to metaphoricisation, as conversion is a relatively productive derivational process in English. The only derivational process that is made use of in one language but ignored in the other is the use of particles, characteristic of English. The reason for that is that only English relies on the use of particles for the formation of verbs, whereas Polish opts for prefixation in this respect. Irrespective of the option favoured by children for coining new words, innovative verbs produced by children are much more iconic than their adult counterparts due to the fact that there is always a causal, not arbitrary, link between the derivative and its base. 3.2.5.5. Conclusion
Inflectional and derivational morphological iconicity in both languages can be accounted for on the grounds of the theory of apperception, according to which human perception has a dual nature. Consequently, considering morphological iconicity, it can be said that in the process of speech perception it is the word stem that children concentrate their attention on. Therefore, in the process of speech production the stem is the part of the word that will always be properly pronounced. Hence, inflectional iconicity consists in the addition of inflectional endings to the word stems, disregarding alternation at the same time. Morphological forms obtained in this way are either thoroughly diagrammatic, or much more diagrammatic than those functioning in the adult language. Considering derivational iconicity, it is again either the word stem that the child focuses on, as in wy-wiązać instead of od-wiązać ‘undo’ from the verb wiązać ‘tie’, or a particular concept out of many concepts that a particular word expresses, as in obcina-czka from obcinać, ‘to cut’, to denote ‘fryzjerka’, ‘hairdresser’. Thus, what is focused on in the process of perception, is automatically highlighted in the process of linguistic production. Consequently, the phenomenon of apperception is responsible for derivational iconicity as well. Analysing the phenomenon of apperception and iconicity in the wider context, it is evident that Rozwadowski’s theory coincides with the cognitive notion of the prototype, discussed and developed by E. Rosch (1978). The presence of prototypical linguistic elements has been stated both in lexical (Southgate and Meints (2000)) and grammatical development (Taylor (1989),
181
Meints et al. (1999)). This means that children first acquire the most basic word meanings and grammatical forms. As far as lexical development is concerned, children narrow down word meanings to the most typical concepts that they express. Therefore, children’s derivational coinages are prototypical in the sense that they express the most typical concepts commonly associated with them. Consequently, these coinages are very often metonymical. When it comes to grammatical development, children first acquire those grammatical forms which are prototypical, i.e. basic and unmarked, for example, word stems: hence the high frequency of noun and verb stems in the speech of Polish and Englishspeaking children. The next subchapter is devoted to the phenomenon of yet another kind of iconicity, namely syntactic iconicity.
3.3. Syntactic iconicity121 As stated in the introduction, iconicity is discernible not only on the morphological but also on the syntactic level of language. In the following subchapter I will analyse the phenomenon of syntactic iconicity on the basis of the utterances from my corpus. I will also prove that child language is much more iconic than adult language as far as its syntax is concerned. First of all, as regards a definition of syntactic iconicity, Brandt (2003) understands syntactic iconicity as syntactic diagrammaticity because of the fact that the syntactic construction diagrams analogically the semantic construal. In other words, the syntactic structure is constructed in such a way that its meaning can be easily and naturally inferred. As far as syntactic iconicity is concerned, it can manifest itself as one of the following principles: I. the principle of quantity; II. the principle of sequentiality; III. the principle of proximity/distance. 3.3.1. The principle of quantity On the basis of my language corpus I can say that the largest number of children’s utterances exhibit syntactic iconicity based on the principle of quantity. Syntactic iconicity seen as the manifestation of the principle of quantity should be understood as reliance on the formula according to which more form 121 The present subchapter has served as the basis of the paper ‘Non-Arbitrary Coding in the Child Language – The Phenomenon of Syntactic Iconicity’ published in 2008 in Language, Literature, Culture and Beyond (pp. 107–113).
182
equals more meaning (Rudzka-Ostyn (1988)). This means that the sentence acquires a new meaning or at least a new shade of meaning once some sentence elements are added to it. For example, the addition of the imperative verb form to an imperative sentence makes the order more emphatic: E (2;0) Sit down, sit down here for ‘Do sit down here’. In my corpus of children’s utterances the addition of new sentence elements consists in the reduplication of the word or a phrase. This process and its implications for the meaning of the whole sentence will be discussed below. First of all, I will focus on the reduplication of the verb or the verb-like form in holophrastic sentences, which has been observed exclusively in the speech of Polish children. The reduplicative forms stand for: 1) repetitive nature of an action: K (1;7–1;10), Z (1;8–2;0) iconic child utterance conventional adult structure płuka-płuka ‘mama płucze naczynia’, i.e. ‘Mummy rinses the dishes’ from the verb płukać, i.e. ‘to rinse’ kopu-kopu ‘kopałam w piasku’, i.e. ‘I was digging in the sand’ from the verb kopać, i.e. ‘to dig’ drapu-drapu ‘drapię się’, i.e. ‘I am scratching myself’ from the verb drapać się, i.e. ‘to scratch’ malu-malu ‘tato malował’, i.e. ‘Daddy was painting’ from the verb malować, i.e. ‘to paint’ myju-myju ‘myjemy się’, i.e. ‘We are washing ourselves’ from the verb myć się, i.e. ‘to wash oneself’ 2) duration of an action: iconic child utterance kręci-kręci kąpu-kąpu fur-fur bo-bo122
conventional adult structure ‘to się wciąż kręci’, i.e. ‘It is still spinning’ from the verb kręcić się, i.e. ‘to spin’ ‘kapałam się długo’, i.e. ‘I was bathing for a long time’ from the verb kąpać się, i.e. ‘to bathe’ ‘ptak frunie’, i.e. ‘The bird is flying’ from the verb frunąć, i.e. ‘to fly’ ‘ciągle boli’, i.e. ‘It is still aching’ from the verb boleć, i.e. ‘to ache’
122 This form, as well as the preceding one, i.e. bo-bo are regarded as verb-like forms because for children they play the role of verbs although they do not function as verbs in adult language.
183
Here reduplication serves the purpose of underlining the intensity of the action, i.e. either its repetition, or lengthy duration. In this way reduplicative verbs are examples of syntactic diagrammatic iconicity. Adult Polish is not iconic as far as the expression of the repetition of an action or its duration is concerned, as these concepts are conveyed by lexical means, i.e. appropriate adverbs like ciągle ‘all the time’, and wciąż ‘still’, as in Ona ciagle/wciąż narzeka ‘She complains/is complaining all the time’ to emphasise duration; or znowu ‘again’, kilka razy ‘a few times’, as in On dzwonił kilka razy ‘He has rung a few times’, to mark the repetition. Thus, the language of children is much more iconic in this respect. Syntactic iconicity based on the principle of quantity manifests itself not only through the reduplication of the verb but also through the reduplication of other parts of the speech as well as whole sentence elements. Unlike verb reduplication, this kind of reduplication can be observed both in my Polish and English data: E (1;8–2;0) iconic structure Girl, girl mixing Big monster, big monster needs pasta pasta’ Eat you, you Baby, I’ll rock my baby in it I’ve been drawing, drawing and drawing long time’ Sit down, sit down here No Mazie, no Mazie Don’t, don’t, don’t Put in door, a door, next door to it He need it, he needs some Jk (2;0) Daddy, daddy work Mummy, mummy book Come mummy, come, come, come, mum 123
conventional adult structure ‘It is123 the girl who is mixing’ ‘It is the big monster that needs ‘It is you that I am going to eat’ ‘It is my baby that I’ll rock in it’ ‘I have been drawing for quite a ‘I really want to sit down here’ ‘There is really no Mazie here’ ‘Don’t you ever do it!’ ‘Put in this door next to it’ ‘He really needs it/some’
‘It is daddy who is at work’ ‘It is really mummy’s book’ ‘Mummy, do come’
The sentence elements that are underlined represent the conventional grammatical or lexical counterparts of iconic structures functioning in child language.
184
Z (1;9) Nie, inny, inny ‘No, another, another’ Ciastko, Zuzia chce ciastko ‘Cake, Zuzia wants a cake’ Patrz, patrz, ucieka krzesło ‘Look, look, the chair is running away’ K (2;1–2;3) Dziobie pingwin dziobie ‘Pecking, the penguin is pecking’ Kaczka, kaczka, popatrz, kaczka siedzi ‘Duck, duck, look, the duck is sitting’ Dużo, dużo, wszystkie, dużo, dużo, tutaj dawać, tutaj dawać. ‘Many, many, all, many, many, give here, give here. Złapał tatę, złapał ‘Caught daddy, caught’ Spadła łupka, spadła ‘Fell, the peel, fell’ Mów to, mów to ‘Say it, say it’ Hałas robię, hałas ‘Noise I’m making, noise’ Gorąca leci woda, taka gorąca leci woda ‘Hot water is pouring, such hot water is pouring’
‘Chcę inny, nie ten’ ‘I want another, not this one’ ‘Ciastko Zuzia chce’ ‘It’s a cake that Zuzia wants’ ‘No, popatrz, krzesło ucieka’ ‘Do have a look! The chair is running away’ ‘Pingwin naprawdę dziobie’ ‘The penguin is really pecking’ ‘Popatrz, to kaczka siedzi’ ‘Look, it is the duck that is sitting’ ‘Daj mi teraz wszystkie jakie masz’ ‘Give me now all the ones you have got’ ‘Naprawdę złapał tatę’ ‘He really caught daddy’ ‘Łupka naprawdę spadła’ ‘The peel did fall’ ‘Bardzo cię proszę, opowiedz mi o tym’ ‘Do tell me about it, please’ ‘Robię duży hałas’ ‘I’m making a lot of noise’ ‘Ciągle leci bardzo gorąca woda’ ‘Very hot water is still pouring’
On the basis of the above examples, it is readily remarked that syntactic reduplication performs numerous functions in sentences: 1) it strengthens the request or order: Don’t, don’t, don’t ‘Don’t do it!’; Come mummy, come, come, come, mum ‘Mummy, do come’; Patrz, patrz, ucieka krzesło ‘Look, look, the chair is running away’, i.e. ‘Look! The chair is running away’; or Mów to, mów to ‘Say it, say it’, i.e. ‘Do tell me about it, please’; 185
2) it emphasises completion of an action: Złapał tatę, złapał ‘Caught daddy, caught’, i.e. ‘He really caught daddy’; or Spadła łupka, spadła ’ ‘Fell, the peel, fell’, i.e. ‘The peel did fall’; 3) it emphasises the duration of an action: I’ve been drawing, drawing and drawing ‘I have been drawing for quite a long time’; or Gorąca leci woda, taka gorąca leci woda ‘Hot water is pouring, such hot water is pouring’, i.e. ‘Very hot water is pouring’; 4) it draws attention to a particular thing – Ciastko, Zuzia chce ciastko ‘Cake, Zuzia wants a cake’, i.e. ‘It’s a cake that Zuzia wants’ – or person – Eat you, you ‘It is you that I am going to eat’ – or activity – Dziobie, pingwin dziobie ‘Pecking, the penguin is pecking’, i.e. ‘The penguin is really pecking’; 5) it emphasises the intensity124 of the phenomenon expressed by the noun: Hałas robię, hałas ‘Noise, I’m making noise’, i.e. ‘I’m making a lot of noise’; 6) it emphasises the intensity of a particular feature: Gorąca leci woda, taka gorąca leci woda ‘Hot water is pouring, such hot water is pouring’, i.e. ‘Very hot water is pouring’. As has been shown, adult language expresses the same functions relying on conventional grammatical structures, or lexical means. Thus: 1) a request is marked both in Polish and in English by the use of the particle (bardzo) proszę ‘please’; 2) the completion of an action is shown by the use of the perfective aspect of the past tense: Łupka spadła ‘The peel fell’; 3) the duration of an action is expressed in English by the use of the progressive aspect, I have been drawing, whereas in Polish it is shown by means of appropriate adverbs, e.g. ciągle ‘still’, as in: Bardzo gorąca woda ciągle leci ‘Very hot water is still pouring’; 4) the effect of putting emphasis on a thing or an action is achieved differently in English and in Polish: a) English uses for that purpose cleft sentences, as in It is daddy who is at work, or auxiliaries, as in Do come, or a word such as really, as in He really needs it. b) Polish relies on contrastive word order, as in Ciastko Zuzia chce [‘Cake Zuzia wants’], or the use of the particle to ‘it’ To kaczka siedzi ‘It is the duck that is sitting’, or the use of the adverb naprawdę ‘really’, as in Naprawdę łupka spadła ‘The peel really fell’; 124
The intensity of the phenomenon or a particular feature has been found to be expressed iconically only in the speech of Polish children.
186
5) the intensity of the phenomenon is expressed by using the adjective duży ‘big’, or its synonyms: Robię duży hałas ‘I’m making a lot of noise’; 6) the intensity of the feature is expressed by means of the modifier bardzo ‘very’: bardzo gorąca woda ‘very hot water’. Thus, it is evident that the language of children is characterised by a high degree of syntactic iconicity based on the principle of quantity. Children show a general tendency to employ syntactic reduplication in order to modify the meaning of the sentence, whereas adult users of language use for that purpose conventional, i.e. arbitrary, means. However, this does not mean that adult language is totally deprived of iconicity dependent on the principle of quantity, as this kind of syntactic iconicity can be observed in colloquial and literary language, when the speaker or the writer wants to be emphatic. For example, in colloquial language it is possible to say He ran, ran, and ran until he got to his car. In this sentence the emphasis has been put on the duration of the activity. The standard, i.e. conventional, version of this sentence would be: ‘He had been running for a long time until he got to his car’. However, in order to emphasise duration of the activity the speaker may want to choose to reduplicate the verb denoting it. When it comes to the use of this aspect of syntactic iconicity in literature, it is especially discernible in poetry. Thus, I would like to illustrate it using as an example an excerpt from a poem written by the Polish poet Leopold Staff (2005) entitled Deszcz jesienny ‘The Autumn Rain’:125 O szyby deszcz dzwoni, deszcz dzwoni126 jesienny I pluszcze jednaki, miarowy, niezmienny, Dżdżu krople padają i tłuką w me okno... Jęk szklany... płacz szklany... a szyby w mgle mokną. I światła szarego blask sączy się senny... O szyby deszcz dzwoni, deszcz dzwoni jesienny [...]. ‘The rain rings against the windowpanes, the autumn rain rings And splatters alike, steady, invariable, The drops of drizzle fall and bang against my windowpane... Glassy groan, glassy wail, and the panes soak in the fog. And seeps the sleepy glitter of grey light... The rain rings against the window panes, the autumn rain rings [...].’127 125
It should be noted that iconic expressions in this poem are a by-product of the attempt to achieve an onomatopoeic effect, that is to imitate the repetitive nature of the rain. 126 In this section all the elements written in bold represent examples of syntactic iconicity. 127 Translation mine.
187
In the above excerpt syntactic reduplication manifests itself by the fourfold repetition of the phrase deszcz dzwoni ‘the rain rings’. Its purpose is to emphasise the persistent and incessant nature of the rain. Another feature of syntactic iconicity that can be observed in this excerpt is the repetition of the attribute szklany ‘glassy’, which is an example of the rule of quantity, according to which more meaning equals more form. Thus, the repetition of this attribute brings the intensity of the rain and the noise that it makes to our attention. Moreover, this poem also exhibits other features of syntactic iconicity. First of all, its very title: Deszcz jesienny [Rain autumn] is iconic, as it is an example of the rule of the sequentiality which determines the order of elements in the phrase or in the sentence. In Polish as well as in English an adjective normally precedes the noun: jesienny deszcz ‘autumn rain’. However, the adjective can be shifted to the end of the phrase in accordance with the rule of so-called temporal proximity (Tabakowska (1998: 68)), which states that the elements to be stressed are put at the end of the sentence or the phrase, as they are ‘closer in time to the listener’ and therefore capable of attracting his attention more effectively. Lastly, iconicity is also discernible in the rhythm, created by the use of short lexemes, separated from one another by numerous punctuation marks, and in this way, constituting a reflection of the rhythmical, i.e. persistent and repetitive, nature of the rain. In this discussion of iconicity in literary language, I touched upon the rule of sequentiality, which I will deal with in the next section. 3.3.2. The principle of sequentiality The iconic principle of sequentiality applies to the order that the phrase or sentence elements are arranged in (Dirven (2001)). This order of sentence elements diagrams analogically the semantic construal, i.e. the meaning of such a sentence can be inferred on the basis of the way its elements have been structured. Consequently, a change in the arrangement of the sentence elements entails acquiring a new shade of meaning. Here is the analysis of the way in which the iconic principle of sequentiality operates based on both my Polish and English data: iconic syntactic structure K (2;0–2;5) Płakać będę [cry I will] Głaskać będę [stroke I will] Sama zostać muszę [alone I must stay]
188
symbolic syntactic structure ‘Będę płakać’, i.e. ‘I wil cry’ ‘Będę głaskać’, i.e. ‘I will stroke’ ‘Muszę zostać sama’, i.e. ‘I must stay alone’
Czekać tutaj lubię [waiting here I like] Składam ja [folding I am]128 Ścieram ja [wiping I am]
‘Lubię tutaj czekać’, i.e. ‘I like waiting here’ ‘Ja składam’, i.e. ‘I am folding’ ‘Ja ścieram’, i.e. ‘I am wiping’
E (1;11–2;4) Dark he is Too many stripes he have got That’s I needs It’s Ella gonna do This Ella drawed
‘He is really/very dark’ ‘He has got far too many stripes’ ‘That is what I need’ ‘Ella is gonna do it (not something else) ‘Ella drew this (not something else)
All the children’s utterances presented above are iconic because the meaning that they convey may be inferred on the basis of the order of their elements. The sentences constructed by children are characterised by a marked word order, i.e. the new information has been shifted to the initial position, whereas the unmarked word order both in Polish and in English is characterised by the new information being supplied in the final position. The change of the order of the sentence elements entails altering the meaning that the sentence expresses because it draws the listener’s attention to the word whose position has been changed. For example, in the sentence Dark he is it is the concept expressed by the adjective that has been emphasised due to the change of its position in the sentence. Similarly, the utterance Płakać będę [cry I will] focuses on the activity of crying because the verb denoting the very activity has been shifted to the beginning of the sentence. Thus, it can be said that the structure of the above sentences is governed by the principle of sequentiality, as the information considered to be the most important occupies the initial place in the structure of the sentence in the same way as it ‘occupies the first place’ in the speaker’s, i.e. the child’s mind. As far as adult language is concerned, it is far less iconic in this respect because changing the word order is not the only method of highlighting particular sentence elements. In contrast, Polish – and especially English – employ either lexical or grammatical means to express emphasis much more often. 3.3.3. The principle of distance According to the iconic principle of distance, those sentence elements that are conceptually close to each other are also close to each other on the linguistic level. The principle of distance is capable of accounting for the occurrence of the 128
The very presence of the personal pronoun ja ‘I’ is iconic as well, as normally, i.e. in adult Polish I would say ścieram ‘I am wiping’ and składam ‘I am folding’.
189
following gender and person concordance that has been observed in my Polish data: children’s iconic language Kolega jest zmęczona Ta kierowca Jaka powodzia?
adult conventional language ‘Kolega jest zmęczony’, i.e. ‘The friend is tired’ ‘Ten kierowca’, i.e. ‘This driver’ ‘Jaka powódź?’, i.e. ‘What flood?’
In the above utterances there exists a close conceptual relationship between the head of the noun phrase and its modifier. This relationship has been reflected on the linguistic level. The conceptual proximity of particular phrase elements has a direct influence on the grammar of the whole phrase. Thus, the inflectional endings of the modifier (ta ‘this’) or the complement (zmęczona ‘tired’) are adjusted to what seems to be the gender of the noun. At this point it should be noted that the noun’s gender has been interpreted here in accordance with the iconic principle of transparency of encoding, according to which a given signans represents one and only one signatum. Thus, children assume that the nominative inflectional morpheme -a is used to denote feminine nouns only and they classify all nouns with the ending -a as feminine. Consequently, they choose feminine modifiers or feminine complements to go with the nouns considered to be feminine. The opposite process can also be observed, namely the noun ending can be influenced by the gender of the noun modifier or the question word. Hence the phrase Jaka powodzia? instead of Jaka powódź? ‘What flood?’ In this case, under the influence of the form of the question word, the child added the -a ending to the noun whose feminine gender has not been overtly marked. The aim of this operation has been the standardisation of inflectional endings for the sake of linguistic proximity resulting from conceptual proximity. Yet another case of the application of the rule of distance would be the following question: K (3;0): Hien-a to dziewczynka, a chłopiec to hien? ‘Hyena (fem.) is a girl and the boy is a hyena (masc.)?’ Here, again, linguistic proximity has been the reflection of conceptual proximity, as the child has assumed that the change into the masculine gender is tantamount to the subtraction of the feminine noun ending -a.
Conclusion Summing up, the language of children is much more iconic than the language of adults as far as its syntactic structure is concerned. As the above examples have shown, child language very often exhibits a similarity between 190
the structure of the sentence and its meaning. The syntactic iconicity of child language manifests itself through three principles: quantity, sequentiality, and distance. As has been shown, it is the principle of quantity that is responsible for the syntactic iconicity of the vast majority of constructions. The principle of sequentiality and the principle of distance are far less common in the language of children. In contrast, the language of adults hardly ever relies on syntactic iconicity, resorting instead to conventional linguistic operations, such as lexical or grammatical means. On the whole, child language is much more iconic than adult language as far as both its syntax and morphology are concerned. Morphological and syntactic iconicity is tantamount to cognitive simplicity and transparency of encoding, which are of primary importance for children. The simpler and the more transparent a given linguistic structure is, the easier it is to acquire. Moreover, the existence of iconicity in child language is a manifestation of its oversimplification for the sake of the clarity of meaning and form.
191
Conclusions
The main aim set to this book has been to account for the process of morphological and syntactic acquisition on the basis of the empirical data. Being aware of the fact that this study is by no means all-exhaustive, it is still possible to indicate certain universal tendencies and draw general conclusions casting some light on the nature of the early stages of linguistic development. Therefore, I would like to propose that the process of language acquisition is shaped by the following three mechanisms: 1. In the initial stages of linguistic development children master stereotypical formulas and fixed phrases, such as there isn’t, I want, no cakes, take that, daddy sleeping, etc., exhibiting little or no linguistic creativity; 2. These fixed and stereotypical linguistic units become consolidated into larger units, without being processed in the slightest degree;129 3. With the passage of time, i.e. as children gradually become capable of abstract thinking, they make numerous generalizations about the structure of language, and, as a result, create their own morphological and syntactic rules that are natural and unmarked; that is, both linguistically and cognitively simple. As far as the early period of language acquisition, i.e. the telegraphic speech, is concerned, it has been discussed in chapter II, and it has been shown that at that stage children produce fixed phrases, with numerous content or functions words deleted from them. These phrases are treated by children as unanalysable and indivisible chunks that can be used irrespective of the situational context. For instance, one of the children studied kept repeating her mother’s commands and orders when making requests or commenting on her own activities, e.g. dam łyżeczkę ‘I will give a teaspoon’ instead of ‘daj łyżeczkę’, i.e. ‘give me a teaspoon’, or nie złam ‘do not break’ instead of ‘nie złamię’, i.e. ‘I will not break’. Another example of treating the phrase as an indivisible whole, characterised by the number of various applications would be the utterance Under table, meant to stand for four communicative intentions: ‘Put it under the table’, ‘It is under the table’, ‘I have put it under 129
192
This hypothesis has been advanced by Tomasello (1998b).
the table’, and ‘I am sitting under the table’. As can be observed, the change of context entails no morphological or syntactic processing. Thus, children do not actively construct their own utterances but master stereotypical phrases, heard in their caretakers’ speech, applying them not only in the context they are meant for but also in other contexts without changing their form. When it comes to the next stage of the language acquisition process, i.e. consolidating simple units into complex units, I would like at this point to provide evidence supporting the claim that more sophisticated linguistic units are formed through the combination of basic linguistic units (/ indicates the boundary between two basic structures): 1. ENGLISH SAMPLE Utterances that are the outcome of consolidation Horsie/ on water Show me/ the shoes Dog/ on the boat Shoes/ too big I find/ no Rosie Ella/ fix it Ella/ suck her thumb Ella/ have a froggie Ice-cube/in the soup? Ella sitting/in your chair There is/two blocks Ella/for daddy Can I do/rock a baby? What’s/he say? I’m/ sew I’m/pull it/in the big whole Where is that/go? Can you see it/chicken here? He take it/that was/crunchy pills Wanna a/have/sip my milk This is/draw a lion I’m an/unbuckle it There’s/no trains There wasn’t/any typhoons
Intended meaning ‘I can see the horsie on the water’ ‘Show me the shoes’ ‘There is a dog on the boat’ ‘The shoes are too big’ ‘I have not found Rosie’ ‘Ella is fixing it’ ‘Ella is sucking her thumb’ ‘Ella has a frog’ ‘Did you put the ice-cube into the soup?’ ‘Ella is sitting in your chair’ ‘There are two blocks’ ‘Ella has made some tea for daddy’ ‘Can I rock my baby?’ ‘What is he saying?’ ‘I’m sewing’ ‘I’m pulling it into the big hole’ ‘Where did that go?’ ‘Can you see the chicken here? ‘He took some crunchy pills’ ‘Do you want to have a sip of my milk?’ ‘I’m going to draw a lion’ ‘I’m going to unbuckle it’ ‘There are no trains’ ‘There weren’t any typhoons’
Even more sophisticated messages are the result of consolidation: 193
Other boot/help me Cut it/small pieces She feed/the bottle 2. POLISH SAMPLE Mamy/rączkę tutaj ‘Mother’s/hand here’ W czajniczku/woda ‘In the kettle/water’ Tam/obrazki dwa ‘There/two pictures’ Brzuszek/nie boli ‘Tummy/does not ache’ Tutaj/brzuszek mamy ‘Here/mummy’s Żaba/piasku ‘Frog/the sand’ Siedzieć/Kinga ‘Sit/Kinga’ Chcę/pani przyszła ‘I want/lady come’ Jak/weź to? ‘How/take it?’ Herbatka ta/pyszna ‘This tea/delicious’ Gdzie/do drugiej cioci? Where/to the other aunt?’ Nie je misiu/to? ‘Does the teddy-bear not eat/this?’ Drugie/nie ugotowałam ‘Another/I have not cooked’ Takie trzeba/bierz ‘This you need/take’
‘Help me to take off the other boot’ ‘You can cut small pieces with it’ ‘She is feeding them with the bottle’
‘Namalujmy tutaj rączkę mamy’ ‘Let us paint mother’s hand here’ ‘W czajniczku jest woda’ ‘There is water in the kettle’ ‘Tam są dwa obrazki’ ‘There are two pictures there’ ‘Nie boli mnie brzuszek’ ‘I have not got a stomach-ache’ ‘Tutaj mama ma brzuszek’ ‘Here is mummy’s tummy’ ‘Żaba wpadła do piasku’ ‘The frog has fallen into the sand’ ‘Kinga chce siedzieć’ ‘Kinga wants to sit’ ‘Chcę żeby pani przyszła’ ‘I want the lady to come’ ‘Jak to wziąć?’ ‘How shall I take it?’ ‘Ta herbatka jest pyszna’ ‘This tea is delicious’ ‘Gdzie jest kartka z telefonem do drugiej cioci?’ ‘Where is the sheet of paper with the phone number of the other aunt?’ ‘Czy misiu tego nie je?’ ‘Does the teddy-bear not eat this?’ ‘Nie ugotowałam drugiego obiadu’ ‘I have not cooked another dinner’ ‘Trzeba takie wziąć’ ‘You need to take these’
As in the English sample, consolidation can result in still more sophisticated structures: 194
To małe/pokaż/gdzie postawić? ‘This little/show/where to put?’ Gdzie położyć/nie wiem ‘Where to put/I don’t know’ Daleko/karteczkę chcę/tam ‘Far away/I want a sheet of paper/there’
‘Pokaż mi gdzie postawić to małe naczynie’ ‘Show me where to put this small dish’ ‘Nie wiem gdzie to położyć’ ‘I do not know where to put it’ ‘Chcę położyć karteczkę tam daleko’ ‘I want to put this sheet of paper far away from me’
The above presented data constitutes just a small sample of all the utterances created as a result of consolidation. As Tomasello (1998b) has remarked, children consolidate not just isolated linguistic categories but entire constructions that they have acquired earlier. These constructions differ from one to another with regard to their internal structure, size and level of abstraction. Evidently, the question may be posed here: what entitles us to adhere to the hypothesis that the emergence of more complicated structures is the result of consolidation? The answer to this question is straightforward: on the basis of the above data it can be easily observed that all the utterances represent no more than a combination of stereotypical formulas acquired by children at the very beginning of their syntactic and morphological development, i.e. during the period of telegraphic speech. For example: Horsie/on water Ella/suck her thumb Ella/sitting/in your chair What’s/he say? Drugie/nie ugotowałam Żaba/piasku
130 131
is a result of the consolidation of the noun horsie and the prepositional phrase on water; is a result of the consolidation of the noun Ella and the verb + direct object suck her thumb; is a result of the consolidation of the noun Ella, present participle sitting and the prepositional phrase in your chair; is a result of the consolidation of the stereotypical formula what’s and pronoun + verb he say;130 ‘Another/I have not cooked’ is a result of the consolidation of the stereotypical formula drugie ‘another’ and negation + verb nie ugotowałam ‘I have not cooked’; ‘Frog/sand’ is a result of the consolidation of the inanimate noun żaba ‘frog’ and the locative noun piasku ‘sand’, etc.131
The pattern of pronoun + verb can be treated as a subtype of the noun + verb pattern. All the patterns of fixed phrases and stereotypical formulas can be found in chapter II.
195
As can be easily observed, the constituents of each utterance, i.e. fixed phrases and stereotypical formulas have been used in all the above cases unprocessed, which means that their form has not been changed in any way to adjust to the new context and to fulfil the requirements of well-formedness of new utterances. Thus, it is undeniable that all the more complex structures are not created from single linguistic units, i.e. lexemes, but from whole language chunks, as each utterance can be easily broken down into two or more fixed phrases or stereotypical formulas that have been acquired earlier. Moreover, the consolidated utterances consist merely of those elements that are cognitively salient, and, consequently, they are deprived of those function and content words that are not important from the child’s point of view. This is a direct consequence of the fact that only cognitively salient elements constitute basic patterns of the early child speech from which all the more sophisticated language units, i.e. consolidated utterances, are made. When it comes to the morphological form of the nouns and verbs used in all the above phrases and utterances, they are not used by children as abstract representations accounting for particular functions of words in a sentence. They should rather be regarded as linguistic units that have been memorised in the form in which they were first heard in the language input. For example, the utterance Jak/weź to? ‘How/take it?’ for ‘Jak to wziąć’, i.e. ‘How to take it?’ contains the imperative form of the verb weź ‘take’ instead of the infinitive ‘wziąć’, i.e. ‘to take’, because weź ‘take’ is the form that the child first heard in requests directed at her, and consequently acquired this form first. Thus, the form of individual lexemes has not been modified in any way, and, consequently, the structure of the fixed phrases and stereotypical formulas remains unchanged. Therefore, these utterances are characterised by morphological divergences from the adult system. However, with the passage of time, children’s grammar begins more and more to resemble adult grammar as children begin to use morphological rules productively. What enables this transition from the acquisition of morphological forms and syntactic structures shaped by rote-learning to the creative application of morphological rules in the process of mastering the grammatical system of the language? As Langacker (2000) has stated, proposing his usage-based model of language acquisition, linguistic units are abstracted from usage events on the basis of those elements that are recurrent and common for all the units. What is meant by this is that children actively construct their linguistic system not from some abstract innate patterns, as generativists have proposed, but by making and verifying hypotheses concerning linguistic structures on the basis of the language input aimed at them. Hereby, relying on my language corpus, I would like to account for the beginnings of morphological processing both in Polish and in English. In chapter 196
I I have presented the hypothesis concerning the nature of the emergence of abstract schemas from the language input that the child is confronted with (Dąbrowska (2003)). In this view, abstract schemas are formed as the child acquires more and more formulas, i.e. specific patterns of language characterised by the same internal structure. This finding has been made by Dąbrowska in the course of studying the acquisition of questions in English. However, I believe that it can also be relied on to explain some aspects of morphological development in Polish and in English. To begin with, Dąbrowska has stated that the process of transition from concrete formulas to abstract schemas takes place as a result of generalisations made about the formulas’ structures. As far as my data and the phenomenon of morphological development are concerned, I would like to propose that in the process of acquiring formulas, children make overgeneralisations concerning their structure, which leads to morphological iconicity.132 In my opinion the transition from the pre-morphological stage to the morphological stage is overgeneralisationbased because it has been proved by many researchers (Brown (1973), Cazden (1968), Kuczaj (1977), Slobin (1971)) that overgeneralisation plays a key role in the process of language acquisition. It has also been stated that the principles governing overgeneralisation are the cognitive and linguistic simplicity of the language units and their frequency in the language input (Slobin (1985)). This means that grammatical rules which are linguistically and cognitively simple and which operate extensively in the language replace those that are linguistically complicated, or irregular, and rarely applied. Thus, formulas that are cognitively and linguistically simple and fairly frequent in the language input constitute the basis for overgeneralising morphological rules established in the language (when talking about morphological development, by formulas I mean single lexemes with inflectional endings attached to them). Let me now discuss the process of the emergence of abstract morphological units from the set of formulas on the basis of acquiring the system of past tense endings by English-speaking children. Several steps leading to the establishment of the internalised mechanism of past tense formation can be differentiated. Firstly, long before realising the existence of morphological rules operating in language, children acquire fixed phrases containing verbs, regarding them as either uninflected or indivisible units, without knowing that the verb form is a variable subject to numerous morphological changes dependent on the communicative intention of the utterance. Thus, the phrase Fix it is regarded by children as a single unit denoting ‘tense- and aspect-unspecified activity performed on the object’ that can be and is used by children in many 132
When it comes to syntactic iconicity, I believe that it is a result of overgeneralisations as well. This issue will be discussed further in this section.
forming
197
communicative contexts.133 The possible intended meanings of the phrase Fix it include the following: ‘Ella wants to fix it’ ‘Ella fixed it’ ‘Ella is going to fix it’ ‘Ella is fixing it’ ‘Will you fix it?’ ‘Can I fix it?’ ‘You should fix it’ ‘Are you fixing it?’ Consequently, the unanalysability of a phrase containing an uninflected verb and the lack of indexicality is tantamount to its ambiguity. Evidently, the more indices, i.e. grammatical morphemes, the phrase contains, the less ambiguous it is. On the other hand, in the initial stages of language development the presence of inflections does not indicate that the utterance or any of its elements can be considered to be analysable and componential from the child’s point of view. In fact, just the opposite is the case, i.e. a phrase containing an inflected verb is as unanalysable as one deprived of grammatical morphemes. The reason for this is that children, going through the initial stage of their linguistic development, first learn entire phrases combined with their communicative intention, without subjecting their constituents to morphological analysis.134 Thus, the acquisition of the phrase Daddy washed it does not proceed according to the rule: ‘use the noun followed by the verb with the -ed ending attached to it and afterwards supply the object’, but is a result of complying with the rule ‘use the whole phrase when you want to communicate what happened’. Consequently, the past form of the verb, i.e. wash-ed is considered to be an unanalysable and indivisible entity. As far as the next stage of the grammatical development is concerned, its onset is marked by progress in the child’s cognitive development, accompanied by an increase in the child’s phonological memory, which leads to the emergence of his ability to categorise linguistic items on the basis of their common phonological features. According to Dąbrowska (2003), this categorisation process begins with the accumulation of formulas sharing the same phonological 133
In chapter II, devoted to the explanation of telegraphic speech, I have presented the extensive applicability of numerous fixed phrases and stereotypical formulas prevalent at the beginning of linguistic development. 134 In the light of the construction grammar approach constructions are the basic units subject to acquisition (Croft (2001)).
198
and grammatical features. When discussing the acquisition of the notion of the past, all the past verb forms – played, phoned, walked, etc. – should be understood as formulas. As these formulas are superimposed on one another, their common feature, i.e. the inflectional ending -ed, is strengthened, which leads to the establishment of a schema, i.e. the abstract representation of past verb forms: VERB + -ed. Since, to the child’s mind, irregular past tense verb forms do not share any common features because their superimposition on one another does not lead to the establishment of an abstract schema, their existence is overlooked at the onset of the period of morphological processing.135 At the same time children begin to create an abstract representation of the regular past verb forms. Consequently, they realise that in order to talk about something that happened in the past it is necessary to add the inflectional morpheme -ed to the verb. This awareness makes them mark irregular verbs for the notion of past in a conspicuous way, i.e. by overregularising them. Thus, the fact that irregular verbs are not morphosemantically transparent and semantically compositional triggers off a process of overgeneralisation, thanks to which the notion of past can be expressed overtly. Such a path of acquisition is in accordance with the assumptions of Natural Morphology, in the light of which unmarked forms are diagrammatic, i.e. both cognitively and linguistically simple, and are therefore given priority in the language acquisition process over irregular, i.e. marked, forms. Thus, as has been proved in chapter III, children apply diagrammatic rules to those processes that are marked, i.e. morphologically irregular. In other words, they overgeneralise the rules of morphological naturalness, which leads to a high degree of iconicity in child language. Hence the great number of overregularised and diagrammatic past tense forms in the speech of children. The above described phenomenon of overregularising grammatical endings, leading to iconicity, proceeds along the same paths for all the morphological processes, both in Polish and English. For example, Polish children overgeneralise the rules of the declensional system and produce a great many iconic nouns, e.g. pies-y instead of psy ‘dogs’ (diagrammatic plural), pantofel-a instead of pantofla ‘slipper(s)’ (diagrammatic genitive), etc. In addition, they simplify the conjugational system of a great many verbs, adjusting it to the principles of Natural Morphology. Hence the numerous cases of iconicity within the verb system as well, e.g. spa-my instead of śpimy (diagrammatic present), wysyspa-my instead of wysypiemy ‘we will empty out’ (diagrammatic future), zryw-ną-ł instead of zerwał ‘he picked’ (diagrammatic momentary aspect), etc. 135
At the pre-morphological stage irregular verbs are learned by rote.
199
In English as in Polish reliance on the oversimplification and overgeneralisation strategy is equally well discernible, although not as prevalent as in Polish, due to the fact that English is characterised by an impoverished morphological system. Thus, apart from the diagrammatic past – take-d for ‘took’, eat-ed for ‘eat’, etc. – there is also the diagrammatic plural: foot-s for ‘feet’, mouse-s for ‘mice’, etc. The same phenomenon, i.e. overgeneralising morphological endings, can be observed in the process of word formation. For example, English-speaking children overgeneralise the agent and instrument suffix -er, creating the following innovative nouns: presser to denote ‘a button that needs to be pressed’, hider to denote ‘a wastepaper basket that can be used for hiding oneself’, reacher to denote ‘a person capable of reaching high’, etc. Thus, the outcome of the overgeneralisation process is again a high degree of iconicity in child speech, manifesting itself in the formation of numerous iconic neologisms. Evidently, if in the construction of their language children strive for simplification, linguistic and cognitive clarity, and unmarkedness, the following question must be posed: why is it the diagram that is so heavily relied on in child language and not the symbol, i.e. the morphological metaphor, which seems even simpler as it does not involve any internal changes within the structure of the word? This question seems even more justified in the light of my observation made in chapter II that child language is prototype-dependent, i.e. that children resort in many cases to the prototypical, i.e. grammatically and cognitively simplest forms, such as infinitives, nominatives, etc. In trying to explain this issue, it must be admited that, on the one hand, prototypicality is unquestionably a feature of child language, especially in its early stages. On the other hand, it should be borne in mind that no natural language is acquired in a vacuum, but in society, among, and from people. As Langacker (2000) and many cognitive linguistics-oriented first language acquisition researchers (Tomasello (2003), Dąbrowska and Kubiński (2003), Goldberg (2003)) have stated, the child’s idiolect is formed from the language input that he is confronted with. Consequently, the child, having come to the stage at which he discovers the rules governing the structure of his language, notices the existence of a great many morphological endings, performing a wide variety of functions. Because of that, he realises that to speak a language means to attach inflectional endings to lexemes, however simplified this process might be. Therefore, with this development, prototype-dependent language, i.e. one relying on the use of morphological metaphors, gradually vanishes. Moreover, the theory of Natural Morphology views a diagram as a very natural sign, as it is both perfectly segmentable and semantically motivated. Therefore, it is heavily relied on, as the more natural, i.e. linguistically and cognitively simple the sign is, the more extensively it is relied on by language-users. 200
Arrival at the stage of processing the language input entails yet another development, namely the emergence of syntactic iconicity in children’s speech. In my opinion, syntactic iconicity is the result of the overgeneralisation of linguistically natural rules responsible for encoding syntactic information. Thus, because, for example, in the language input important information, which is cognitively and linguistically simple, i.e. natural, is very often repeated children form the hypothesis that important information is always highlighted in the process of repetition. For instance, caretakers often say twice or three times those parts of the utterance they consider important and that they would like their children to focus on, e.g. Don’t do that. Don’t do that!, or Come here. Come here, at once! As a result, children begin to form sentences in which the elements they want their listeners to concentrate on are repeated, e.g. Big monster, big monster needs pasta for ‘It is a big monster that needs pasta’, Eat you, you for ‘It’s you that I am going to eat’, or Girl, girl mixing for ‘It is a girl who is mixing’. I am of the opinion that the formulas containing the sentence elements that are repeated are again superimposed on one another, and, as a result, abstract schemas emerge. For example, consider the situation in which the child has acquired several utterances consisting of the agent and the repeated verb: Ella mixing, mixing; Ella drinking, drinking; Ella sitting, sitting, etc. These utterances are formulas possessing an identical structure, as they are composed of an agent and the repeated form of the verb. In the end, superimposed on one another, they lead to the emergence of the following abstract schema: PERSON + VERB + VERB meant to emphasise repetition, completion or performance of the activity. Thus, it can be claimed that not only morphological, but also syntactic iconicity is the result of formulas being superimposed on one another, and overgeneralising the rules concerning the use of abstract schemas derived from them. When discussing the question of the linguistic sign in child language, the issue of indexicality cannot be passed over – all the more so because iconicity and indexicality are considered to be the two most important tendencies that shape each linguistic system. However, on the basis of my language corpus the strong dominance of one tendency over the other can be noticed; namely, it can be stated that the language of children is more iconic than the language of adults but less indexical. Hereby, I would like to account for this phenomenon on the grounds of the theory of Natural Morphology. According to Galeas (1998), an index is the simplest sign but it is less natural than an icon because of the lack of analogy between signans and signatum. Thus, viewing the issue of the linguistic sign in the light of the theory of Natural Morphology, which proposes that naturalness is the key principle in language, it comes as no 201
surprise that the icon is the dominant feature of child language. Therefore, the language of children abounds in a great many iconic structures as far as its morphological and syntactic structures are concerned, and, at the same time, it is deprived of many grammatical morphemes obligatory in adult language, especially as they do not substantially modify the lexical signatum. On the whole, the presence of an icon in child language is more noticeable than the presence of an index. To recapitulate, I would like to remark that all the above findings have been made on the basis of my own language corpus of the three Polish and three English-speaking children. Although the three-child language sample is said to be representative enough to carry out observations of the first language acquisition process (Ingram (1989)), I realise that the statements presented in the foregoing require further investigation based on a more extensive language sample in order to confirm the hypotheses concerning the issues, such as prototypicality, unmarkedness, consolidation of formulas and the emergence of abstract linguistic schemas. This area of research would be quite interesting to follow, particularly with regard to the acquisition of Polish, which has been far less extensively studied than the acquisition of English, especially within the cognitive framework. In addition, for the sake of the universality of the above findings, it would seem essential to carry out more cross-linguistic research on the above discussed topics; that is, to broaden the spectrum of languages to be analysed, and to rely on languages other than Polish and English.
202
References
1. Aitchison, J. (1989) The Articulate Mammal. London: Unwin Hyman. 2. Anderson, J.R. (1993) Rules of the Mind. Hillsdale, N.J.: Lawrence Erlbaum Associates. 3. Barret, M.D. (1995) “Early Lexical Development” in P. Fletcher and B. MacWhiney (eds). 362–392. 4. Bates, E. (1976) Language and Context. New York: Academic Press. 5. Bates, E., P.S. Dale and D. Thal (1995) “Individual Differences and their Implications for Theories of Language Development” in P. Fletcher and B. MacWhiney (eds), 96–151. 6. Bates, E. and B. Mac Whiney (1978) “Sentential Devices for Conveying Giveness and Newness: a Cross-Cultural Developmental Study” in Journal of Verbal Learning and Verbal Behaviour 17, 539–558. 7. Bąk, P. (1989) Gramatyka języka polskiego. Warszawa: Wiedza Powszechna. 8. Bellugi, U. and R. Brown (eds) (1964) The Acquisition of Language. Monographs of the Society for Research in Child Development 29. 9. Bellugi, U., K. Van Hoek, D. Lillo-Martin, and L. O’Grady (1993) “The Acquisition of Syntax in Young Deaf Singers” in D.Bishop, D. and K. Mogford (eds) Language Development in Exceptional Circumstances. Hillsdale, N.J.: Lawrence Erlbaum Associates, 132–50. 10. Berko, J. (1958) “The Child’s Learning of English Morphology” in Word 14, 150–177. 11. Bickerton, D. (1990) Language and Species. Chicago: University of Chicago Press. 12. Blasdell, R. and P. Jensen (1970) “Stress and Word Position as Determinants of Imitation in in First-Language Learners” in Jornal of Speech and Hearing Research 13, 193–202. 13. Blank, A. and P. Koch (eds) (1999) Historical Semantics and Cognition. Berlin: Mouton de Gruyter 14. Bloom, L. (1970) Language Development: Form and Function in Emerging Grammars. Cambridge, Mass.: the MIT Press. 15. Bloom, L. (1971) “Why not Pivot Grammar?” in Journal of Speech and Hearing Disorders 36, 40–50. 16. Bloom, L. (1973) One Word at a Time: The Use of Single Word Utterances before Syntax. The Hague: Mouton. 17. Bloom, L., L. Lightbown and L. Hood (1975) “Structure and Variation in Child Language” in Monographs for the Society for Research in Child Development 40 (2, Serial No. 160). 18. Bloomfield, I. (1933) Language. New York: Holt, Rinehart and Winston. 19. Blount, B. (1969) Acquistion of Language by Luo Children. Unpublished doctoral book. Berkeley, CA: University of California. 20. Borer, H. and K. Wexler (1987) “The Maturation of Syntax” in Roeper, T. and E. Williams (eds) Parameter Setting and Language Acquisition. Dordrecht: D. Reidel, 128–151. 21. Bowerman, M. (1973) Early Syntactic Development: a Cross-linguistic Study with Special Reference to Finnish. Cambridge: Cambridge University Press. 22. Bowerman, M. (1975) “Commentary on ‘Structure and Variation in Child Language’ by L. Bloom, P. Lightbown and L. Hood” in Monographs of the Society for Research in Child Development 40, 260–278.
203
23. Bowerman, M. (2003) ”Rola predyspozycji kognitywnych w przyswajaniu systemu semantycznego” in E. Dąbrowska and W. Kubiński (eds) (2003), 254–313. 24. Braine, M. (1963) “On Learning the Grammatical Order of Words” in Psychological Review 70, 323–348. 25. Braine, M. (1971) “The Acquisition of Language in Infant and Child” in C. Reed (ed.) The Learning of Language. New York: Appleton, 7–95. 26. Brandt, W. (2003) “Humour and Meaning Construction in Everyday Speech: A Mental Space Analysis”. http://www.hum.au.dk/semiotics. 27. Brenstiern-Pfanhauser, S. (1930) “Rozwój mowy dziecka” in Prace Filologiczne 15, 273–356. 28. Brown, R. (1970) “The First Sentences of Child and Chimpanzee” in R. Brown Psycholinguistics. New York: The Free Press, 208–231. 29. Brown, R. (1973) A First Language: The Early Stages. Cambridge MA: Harvard University Press. 30. Brown, R. and U. Bellugi (1964) “Three Processes in the Acquisition of Syntax” in Harvard Educational Review 34, 133–151. 31. Brown, R. and C. Frazer (1963) “The Acquistion of Syntax” in Ch. Coffer and B. Musgrave (eds) Verbal Behaviour and Learning: Problems and Processes. New York: McGraw-Hill, 158–201. 32. Brown, R. and C. Frazer (1964) “Explorations in Grammar Evaluation” in U. Bellugi and R. Brown (eds), 79–92. 33. Butterworth, B. and M. Harris (1994) Principles of Developmental Psychology. Hove: Lawrence Erlbaum. 34. Bybee, J.L. (1985) Morphology: A Study of the Relation Between Meaning and Form. Amsterdam: Benjamins. 35. Campbell, J. (1982) Grammatical Man, Information, Entropy, Language, and Life. New York: Simon and Schuster. 36. Campbell, R. and Wales, R. (1970) “The Study of Language Acquisition” in J. Lyons (ed.) New Horizons in Linguistics. London: Penguin, 243–269. 37. Caplan, D. (1987) Neurolonguistics and Linguistic Aphasiology: an Introduction. Cambridge: Cambridge University Press. 38. Caplan, D. (1992) Language: Structure, Processing and Disorders. Cambridge MA: the MIT Press. 39. Cazden, C. (1968) “The Acquisition of Noun and Verb Inflections“ in Child Development 39, 433–448. 40. Cazden, C. (1972) Child Language and Education. New York: Holt, Rinehart and Winston. 41. Chini, M. and Galeas, G.C. (1985) “Verbalparadigmen, Prototypizität und Natürliche Morphologie: Überlegungen am Beispiel des Erstspracherwerbs des Italienischen“ in Quaderni Patavini di Linguistica 14, 73–110. 42. Chmura-Klekotowa, M. (1967) ”Neologizmy słowotwórcze w mowie dzieci” in Poradnik Językowy 10, 433–435. 43. Chmura-Klekotowa, M. (1971) “Neologizmy słowotwórcze w mowie dzieci” in Prace Filologiczne 21, 99–235. 44. Chomsky, N. (1957) “A Review of B. Skinner’s Verbal Behaviour” in Language 35, 26–58. 45. Chomsky, N. (1969) Aspects of the Theory of Syntax. Cambridge, Mass.: the MIT Press. 46. Chomsky, N. (1975) Reflections on Language. New York: Pantheon. 47. Chomsky, N. (1981) Lectures on Government and Binding. Dordrecht: Foris Publications. 48. Chomsky, N. (1986a) Knowledge of Language, its Nature, Origin and Use. New York: Praeger.
204
49. Chomsky, N. (1986b) Barriers. Cambridge Mass: the MIT Press. 50. Chomsky, N. (1987) Language in a Psychological Setting. Tokyo: Sophia linguistica. 51. Chomsky, N. (1991) ”Linguistics and Cognitive Science: Problems and Mysteries” in A. Kasher (ed.) The Chomskyan Turn. Oxford: Blackwell, 43–74. 52. Chomsky, N. (1995) The Minimalist Programme. Cambridge Mass: the MIT Press. 53. Clahsen, H. (1990) “Constraints on Parameter Setting: a Grammatical Analysis of some Acquistion Stages in German Child Language” in Language Acquistion 1, 361–91. 54. Clark, E. (1987) “The Principle of Contrast: A Constraint on Language Acquisition” in B. Mac Whiney (ed.) Mechanisms of Language Acquisition. Hillsdale, NJ: Lawrence Erlbaum Associates, 1–33. 55. Clark, E. (1993) The Lexicon in Acquisition. Cambridge: Cambridge University Press. 56. Clark, E. (1996) “Early Verbs, Event Types and Inflections” in C. Johnson and J. Gilbert (eds.) Children’s Language, vol. 9. Mahwah, NJ: Erlbaum, 134–187. 57. Clark, R. (1974) “Performing without Competence” in Journal of Child Language 1, 1–10. 58. Cook, V. (1988) Chomsky’s Universal Grammar: an Introduction. Oxford: Blackwell. 59. Cook, V. (1997) Inside Language. London: Arnold. 60. Cook, V. and M. Newson (1996) Chomsky’s Universal Grammar: An Introduction. Oxford: Blackwell. 61. Crain, S. and K. Wexler (1999) “Methodology in the Study of Language Acquisition: A Modular Approach” in W.C. Richie and T.K. Bhatia (eds) Handbook of Child Language Acquisition. San Diego: Academic Press, 387–425. 62. Croft, W. (1991) Syntactic Categories and Grammatical Relations: The Cognitive Organization of Information. Chicago: The University of Chicago Press. 63. Croft, W. (2001) Radical Construction Grammar: Syntactic Theory in Typological Perspective. Oxford: Oxford University Press. 64. Cutler, A., Hawkin, J.A. and Gilligan, C. (1985) “The Suffixing Preference: A Processing Explanation” in Linguistics 23, 723–58. 65. Darwin, C. (1877) “A Biographical Sketch of an Infant“ in Mind 2, 285–294. 66. Dąbrowska, E. (2000a) “Could a Chomskyan Child learn Polish? The Logical Argument for Learnability” in M.R. Perkins and S. Howard (eds) New Directions in Language Development and Disorders. New York: Plenum, 85–96. 67. Dąbrowska, E. (2000b) “From Formula to Schema: The Acquisition of English Questions” in Cognitive Linguistics 11, 1–20. 68. Dąbrowska, E. (2003) “Od formuły do schematu: akwizycja pytań w języku angielskim” in Dąbrowska, E. and Kubiński, W. (eds). 69. Dąbrowska, E. and Kubiński, W. (eds) (2003) Akwizycja języka w świetle językoznawstwa kognitywnego. Kraków: Universitas 70. De Saussure, F. (2000) Kurs językoznawstwa ogólnego. Warszawa: Wydawnictwo Naukowe PWN. 71. de Villiers, J. and P. de Villiers (1973a) “A Cross-sectional Study of the Acquisition of Grammatical Morphemes in Child Speech” in Journal of Psycholinguistic Research 2, 267– 278. 72. de Villers, J. and P. de Villers (1985) “The Acquisition of English” in D. Slobin (ed.) (1985), 27–140. 73. Dirven, R. (1999) Cognitive English Grammar. Amsterdam: Benjamins. 74. Dirven, R. and G. Radden (2001) “Kognitywne podstawy języka: język i myśl” in E. Tabakowska (ed.) (2001), 15–42. 75. Dresher, B. (1981) “On the Learnability of Abstract Phonology” in C. Baker and J. McCarthy (eds) The Logical Problem of Language Acquisition. Cambridge Mass: the MIT Press, 188–210.
205
76. Dressler, W. (1985a) “On the Predictivness of Natural Morphology” in Journal of Linguistics 21, 321–37. 77. Dressler, W. (1985b) “Suppletion in Word-formation” in J. Fisiak (ed.) Historical Semantics, Historical Word-formation. Berlin: Mouton de Gruyter, 97–112. 78. Dressler, W. and Denes, G. (1988) “Word-formation in Italian-speaking Wernicke’s and Broca’s Aphasics” in W. Dressler and J. Stark (1988) Linguistic Analysis of Aphasic Language. Berlin: Springer, 109–141. 79. Dressler, W. and Stark J. (1988) Linguisitic Analysis of Aphasic Language. Berlin: Springer. 80. Dressler, W., Mayerthaler W., Panagl, O. and Wurzel, W. (1987) Leitmotifs in Natural Morphology. Amsterdam: John Benjamins. 81. Dulay, H., Burt M., and Krashen S. (1982) Language Two. New York: Oxford University Press. 82. Dziubalska-Kołaczyk, K. (1997) “Pre- and Proto- in Polish Phonology and Morphology and their Interrelations” in Poznań Studies in Contemporary Linguistics 33, 158–171. 83. Ellis, N. (1994) Implicit and Explicit Learning of Languages. New York: Academic Press. 84. Ellis, N. (1996b) “Analyzing Language Sequence in the Sequence of Language Acquisition: Some Comments on Major and Ioup” in Studies in Second Language Acquisition 19, 145– 171. 85. Ellis, N. and R. Schmidt (1997) “Morphology and Longer Distance Dependencies: Laboratory Research Illuminating the A in SLA” in Studies in Second Language Acquisition 18, 361–8. 86. Ervin, S. (1964) “Imitation and Structural Change in Children’s Language” in E.H. Lenneberg (ed.) New Directions in the Study of Language. Cambridge, MA: the MIT Press, 163–189. 87. Fillmore, C.J. (1968) “The Case for Case” in E. Bach and R.T. Harms (eds) Universals in Linguistic Theory. New York: Holt, Rinehart and Wiston, 1–87. 88. Fillmore, C.J. (1988) “The Mechanisms of Construction Grammar” in Proceedings of the Annual Meeting of the Berkeley Linguistics Society 14, 35–55. 89. Fillmore, C.J. (1999) “Grammatical Constructions and Linguistic Generalisations: the What’s X Doing Y Construction” in Language 75, 1–33. 90. Fillmore, C.J. (1985) “Syntactic Intrusions and the Notion of Grammatical Construction” in Berkely Linguistic Society 11, 73–86. 91. Fisiak, J. (1978) An Introductory Polish-English Contrastive Grammar. Warszawa: PWN. 92. Fisiak, J. (ed.) (1985) Historical Semantics, Historical Word-formation. Berlin: Mouton de Gruyter. 93. Fletcher, P. and B. MacWhiney (eds) (1995) The Handbook of Child Language. Oxford: Blackwell. 94. Fodor, J. A.,T. Bever and M. Garret (1974) The Psychology of Language. New York: McGraw Hill. 95. Galeas, G.C. (1998) The Parameters of Natural Morphology. Padova: Unipress. 96. Gathercole, V., E. Sebastian and P. Soto (1999) “The Early Acquisition of Spanish Verbal Morphology: Across-the-board or Piecemeal Knowledge?” International Journal of Bilingualism 3, 43–67. 97. Geeraerts, D. (1985) Paradigm and Paradox: Explorations into a Paradigmatic Theory of Meaning and its Epistemological Background. Leuven: Leuven University Press. 98. Gillette, M. and C. Witt (1998) “What is Linguistic Redundancy?” http://gatton.uchicago.edu/ ~wit/redundan.html. 99. Givon, T. (1985) “Function, Structure and Language Acquisition“ in Slobin, D. (ed.) (1985) Vol.2., 1005–1027.
206
100. Gleason, H. (1961) An Introduction to Descriptive Linguistics. Rev. ed. New York: Holt, Rineheart and Winston. 101. Goldberg, A.E. (1995) Constructions: A Construction Grammar Approach to Argument Structure. Chicago: University of Chicago Press. 102. Goldberg, A.E. (1999) “The Emergence of the Semantics of Argument Structure Constructions“ in B. Mac Whiney (ed.) The Emergence of Language. Mahwah, NJ: Lawrence Erlbaum Associates, 197–212. 103. Goldberg, A.E. (2003) “Kształtowanie się semantyki konstrukcji struktury argumentowej w języku angielskim“ in E. Dąbrowska and W. Kubiński (eds). 104. Goldberg, G. (1985) ”Supplementary Motor Area Structure and Function: Review and Hypothesis” in Behavioral and Brain Sciences 8, 567–615. 105. Goldfield, B.A. and C.E. Snow (1989) “Individual Differences in Language Acquisition” in J. Gleason (ed.) The Development of Language. Columbus: Merrill, 307–330. 106. Golinkoff, R.M. and Kerr, L.J. (1978) “Infants’ Perceptions of Semantically Defined Action Role Changes in Filmed Events” in Merrill Palmer Qarterly 24, 53–61. 107. Gopnik, M. and M. Crago (1991) “Familial Aggregation of a Developmental Language Disorder” in Cognition 39, 1–50. 108. Gregoire, A. (1937) L’Apprentissage du Langage: Les Deux Premieres Annees. Paris: Librairie E.Droz. 109. Grzegorczykowa, R., Laskowski R. and H. Wróbel (eds) (1998) Gramatyka współczesnego języka polskiego. Morfologia. Warszawa: PWN. 110. Gussman, E. (1980) Introduction to Phonological Analysis. Warszawa: PWN. 111. Harris, M. and M. Coltheart (1986) Language Processing in Children and Adults: An Introduction. London: Routledge. 112. Hawkins, J. (1983) Word Order Universals. New York: Academic Press. 113. Heine, B. (1997) Cognitive Foundations of Grammar. New York and Oxford: Oxford University Press. 114. Hocket, Ch. (1958) A Course in Modern Linguistics. New York: Macmillan. 115. Hollingdale, R.J. (1999) Nietzsche: The Man and His Philosophy. Cambridge: Cambridge University Press. 116. Howatt, A. (1988) “From Structural to Communicative” in Annual Review of Applied Linguistics 8, 14–29. 117. Hunnicut, S. (1985) “Intelligibility versus Redundancy – Conditions of Dependency” in Language and Speech, vol. 28(1), 47–56. 118. Hyams, N. (1992) “The Genesis of Clausal Structure” in J.M. Meisel (ed.) The Acquistion of Verb Placement: Functional Categories and V2 Phenomena in Language Development. Dordrecht: Kluwer, 215–238. 119. Hyams, N. (1994) “V2 Null Arguments and COMP Projections” in T. Hoekstra and B.D. Schwartz (eds) Language Acquistion Studies in Generative Grammar. Amsterdam: John Benjamins, 140–156. 120. Hyams, N. (1996) “The Underspecification of Functional Categories in Early Grammar” in H. Clahsen (ed.) Generative Perspectives on Language Acquisition. Amsterdam: John Benjamins. 121. Ingram, D. (1974) “The Relationship between Comprehension and Production” in R. Schiefelbusch and L. Loyd (eds), 3–27. 122. Ingram, D. (1975) “Cerebral Speech Lateralisation in Young Children” in Neuropsychologia 13, 103–105.
207
123. Ingram, D. (1978) “Sensori-motor Intelligence and Language Development” in A. Lock (ed.) Action, Gesture and Symbol: the Emergence of Language. New York: Academic Press, 261–290. 124. Ingram, D. (1985) “On Children’s Homonyms” in Journal of Child Language 12, 671–680. 125. Ingram, D. (1989) First Language Acquisition: Method, Description and Explanation. Cambridge: Cambridge University Press. 126. Ingram, D. (1992) “Early Phonological Acquisition: a Cross-linguistic Perspective” in C. Ferguson, L. Menn and C. Stoel-Gammon (eds) Phonological Development: Model, Research, Implications. Timonium, MD: York Press, 423–35. 127. Ingram, D. (1995) “The Cultural Basis of Prosodic Modifications to Infants and Children: a Response to Fernald’s Universalist Theory” in Journal of Child Language 22, 223–233. 128. Jackendoff, R.S. (1977) X Syntax: a Study of Phrase Structure. Cambridge, MA: the MIT Press. 129. Jenkins, J. and D. Palermo (1964) “Mediation Processes and the Acquisition of Linguistic Structure” in U. Bellugi and R. Brown (eds), 141–169. 130. Jespersen, O. (1965) A Modern English Grammar on Historical Principles. Part IV: Morphology. Copenhagen: Ejnar Munksgaard. 131. Jespersen, O. (1971) The Philosophy of Grammar. Paris: Editions de Minuit. 132. Johnson, C. and J. Gilbert (eds.) (1996) Children’s Language, vol. 9. Mahwah, NJ: Erlbaum. 133. Johnston, J.R and D. Slobin (1979) “The Development of Locative Expressions in English, Italian, Serbo-Croatian and Turkish” in Journal of Child Language 6, 529–545. 134. Johnston, J.R. (1985) “Cognitive Prerequisites: the Evidence from Children Learning English” in D. Slobin (ed.) (1985). 961–1004. 135. Kardela, H. (ed.) (1994) Podstawy gramatyki kognitywnej. Warszawa: PWN. 136. Katz, N., J. Baker and J. MacMacnamara (1974) “What is in a Name? A Study of how Children Learn Common and Proper Names” in Child Development 45, 468–473. 137. Kernan, K. (1969) The Acquisition of Language by Samoan Children. Unpublished doctoral book. Berkeley, CA: University of California. 138. Klima, E.S. and Bellugi, U. (1966) “Syntactic Regularities in the Speech of Children“ in J. Lyons and R.Wales (eds) Psycholinguistic Papers. Edinburgh: Edinburgh University Press, 183–207. 139. Klimkowski, K. (2005). “Apperception in Compound Formation as a Manifestation of Iconicity” http://home.hum. Uva.nl/iconicity/conferences. 140. Konieczna E. 2002. “Derivational Neologisms in Children’s Speech from Polish and English Data’ in Studia Anglica resioviensia 1, 52–64. 141. Konieczna, E. 2008. “Non-Arbitrary Coding in the Child Language – The Phenomenon of Syntactic Iconicity” in Language, Literature, Culture and Beyond R. Kiełtyka, D. Osuchowska, E. Rokosz-Piejko. (eds) Rzeszów: Wydawnictwo Uniwersytetu Rzeszowskiego, 107–113. 142. Konieczna, E. (2011a). “Inflectional Diagrammaticity in Polish: A Case Study of Polish Children” in Word Structure 4(1). L. Bauer (ed.) Edinburgh: Edinburgh University Press, 20–53. 143. Konieczna, E. (2011b). “The Role of Morphological Naturalness in the Production of Innovative Verbs in English and Polish: a Comparative Study” in vol. 8, No. 1 SKASE Journal of Theoretical Linguistics, 2–22. 144. Konieczna, E. and G. A. Kleparski (2006). “Motivation in Language: The Case of Metonymically Motivated Innovative Nouns in Children's Speech” in Odisea: Revista de Estudios Ingleses, 103–118. 145. Kuczaj, S.A. (1977) “The Acquisition of Regular and Irregular Past Tense Forms” in Journal of Verbal Learning and Verbal Behaviour 16, 589–600.
208
146. Kuczaj, S.A. (1982) Language Development: Problems, Theories and Controversies. Vol. 1: Syntax and Semantics. Hillsdale, N.J.: Lawrence Erlbaum Associates. 147. Lakoff, G. (1987) Women, Fire and Dangerous Things: What Categories Reveal about the Mind. Chicago: University of Chicago Press. 148. Lakoff, G. (1997) “How Unconscious Metaphorical Thought Shapes Dreams” in D. J. Stein (ed.) Cognitive Science and the Unconscious. Washington, DC: American Psychiatric Press, 89–119. 149. Lakoff, G. and M.H. Johnson (1980) Metaphors I Live By. Chicago: University of Chicago Press. 150. Lakoff, G. and M.H. Johnson (1999) Philosophy in the Flesh: the Embodied Mind and its Challenge to Western Thought. New York: Basic Books. 151. Langacker, R. (2001) “Kotwiczenie, kodowanie, dyskurs” in W. Kubiński and D. Stanulewicz (eds) Językoznawstwo Kognitywne ii. Zjawiska pragmatyczne. Gdańsk: Wydawnictwo Uniwersytetu Gdańskiego, 22–69. 152. Langacker, R. (1997) “The Contextual Basis of Cognitive Semantics” in J. Nuyts and E. Pederson (eds) Language and Conceptualisation. Cambridge: Cambridge University Press, 229–253. 153. Langacker, R. (1988) “A Usage-based Model“ in B. Rudzka-Ostyn (ed.), 127–161. 154. Langacker, R. (2000) Grammar and Conceptualisation. Berlin, New York: Mouton de Gruyter. 155. Langacker, R.W. (1987) Foundations of Cognitive Grammar. Vol. 1. Theoretical Prerequisites. Stanford: Stanford University Press. 156. Langacker, R.W. (1991) Foundations of Cognitive Grammar. Vol. 2. Descriptive Application. Stanford: Stanford University Press. 157. Langacker, R.W. (2003) “Model dynamiczny oparty na uzusie językowym” in E. Dąbrowska and W. Kubiński (eds), 30–117. 158. Lebeaux, D. (1998) Language Acquistion and the Form of Grammar. Unpublished PhD book, University of Massachusetts. 159. Lenneberg, E.H. (1967) Biological Foundations of Language. New York: Wiley. 160. Lenneberg, E.H. (ed.) (1964) New Directions in the Study of Language. Cambridge, MA: the MIT Press. 161. Leonard, L.B. (1976) Meaning in Child Language. New York: Grune & Stratton. 162. Leonard, L.B. (1998) “Grammatical Morphology and the Role of Weak Syllables in the Speech of Italian-speaking Children with Specific Language Impairment” in Journal of Speech, Language and Hearing Research 41, 1363–1374. 163. Lewandowska-Tomaszczyk, B. (1988) Ways to Language. Łódź: Uniwersytet Łódzki. 164. Lieven, E. (1978) “Turn-taking and Pragmatics: Two Issues in Early Child Language Acquisition” in R. Campbell and P.T. Smith (eds) Recent Advances in the Psychology of Language: Language Development and Mother-child Interaction. New York: Plenum Press, 15–23. 165. Lieven, E., J.M. Pine and G. Baldwin (1997) “Lexically-based Learning and Early Grammatical Developments” in Journal of Child Language 24, 187–220. 166. Lightfoot, D. (1989) “The Child’s Trigger Experience: Degree-0 Learnability” in Behavioral and Brain Sciences 12, 321–375. 167. MacWhiney, B. (1982) “Basic Syntactic Processes” in S. Kuczaj (ed.) Language Development: Problems, Theories and Controversies. vol.1.: Syntax and Semantics. Hillsdale, N.J.: Lawrence Erlbaum Associates, 73–136. 168. Mac Whiney, B. (1985) “Hungarian Language Acquisition as an Exemplification of a General Model of Grammatical Development” in D. Slobin (ed.) (1985), 1069–1155.
209
169. Mac Whiney, B. (ed.) (1987) Mechanisms of Language Acquisition. Hillsdale, NJ: Lawrence Erlbaum Associates. 170. Mac Whiney, B. and J. Leinbach (1991) “Implementations are not Conceptualisations: Revising the Verb Learning Model” in Cognition 29, 198–256. 171. MacClelland, J. and D. Rumelhart (eds) (1986) Parallel Distributed Processing: Explorations in the Microstructure of Cognition. Cambridge, MA: the MIT Press. 172. Majewicz, A.F. (1982) ”Understanding Aspect” in Lingua Posnaniensis 24, 29–62. 173. Maratsos, M.P. (1984) “The Acquisition of Syntax” in J. Flavell and E. Markman (eds) Carmichaels’s Handbook of Child Psychology, vol.3, New York: Wiley, 134–156. 174. Maratsos, M.P. and M.A.Chalkey (1980) “The Internal Language of Children’s Syntax: the Ontogenesis and Representation of Syntactic Categories” in K. Nelson (ed.) Children’s Language. New York: Gardner Press, 127–214. 175. Marchand, H. (1969) The Categories and Types of Present-Day English Word-Formation. Munich: C.H. Beck. 176. Marchman, V.A., E.Bates, A.Burkardt and A.B. Good (1991) “Functional Constraints of the Acquisition of the Passive: Toward a Model of the Competence to Perform” in First Language 11, 65–92. 177. Mayerthaler, W. (1981) Morphologische Natürlichkeit. Wiesbaden: Athenaion. 178. Mc Laughlin, B. and J. L.Heredia (1996) “Information-processing Approaches to Research on Second Language Acquisition and Use” in W.C. Ritchie and T.K. Bhatia (eds) Handbook of Second Language Acquisition. San Diego: Academic Press, 213–28. 179. Mc Neill, D. (1966a) “Developmental Psycholinguistics” in Smith, F. and G. Miller (eds) The Genesis of Language: A Psycholinguistic Approach. Cambridge, MA: the MIT Press, 15–84. 180. Mc Neill, D. (1966b) “The Creation of Language by Children” in Lyons J. and R. Wales (eds) Psycholinguistic Papers. Edinburgh: Edinburgh University Press, 99–115. 181. Mc Neill, D. (1970a) “Language before Symbols: Very Early Children’s Grammar” in Interchange 1. 182. Mc Neill, D. (1970b) The Acquistion of Language: The Study of Developmental Psycholinguistics. New York: Harper and Row. 183. Mc Neill, D. (1971) “The Capacity for Grammatical Development in Children” in D. Slobin (ed.), 17–40. 184. Meints, K., K. Plunkett and P.L. Harris (1999) “When does an Ostrich Become a Bird: The Role of Prototypes in Early Word Comprehension” in Developmental Psychology 35, 1072– 1078. 185. Menyuk, P. (1969) Sentences Children Use. Cambridge, Mass: the MIT Press. 186. Miller, G. and D. Mc Neill (1969) “Psycholinguistics” in G. Lindzey and E. Aronson (eds) The Handbook of Social Psychology. Reading, Mass.: Addison-Wesley vol.3, 669–794. 187. Mitchel, R. (1998) Second Language Learning Theories. London: Arnold. 188. Moerk, E.L. (1975) “Piaget’s Research as Applied in the Explanation of Language Development” in Merill-Palmer Quarterly 21, 151–170. 189. Morehead, D.M. and A. Morehead (1974) “From Signal to Sign: A Piagetian View of Thought and Language during the First Two Years” in R.L. Schiefelbusch and L.Loyd (eds). 190. Mowrer, O. (1960) Learning Theory and Symbolic Processes. New York: Wiley and Sons. 191. Nagórko, A. (1998) Zarys gramatyki języka polskiego. Warszawa: PWN. 192. Nelson, K. (1975) “Individual Differences in Early Semantic and Syntactic Development” in Annals of the New York Academy of Sciences 263, 132–139. 193. Nelson, K. (1981) ”Individual differences in language development: implications for development and language” in Developmental Psychology 17, 170–187.
210
194. Nelson, K. (1975) “Individual differences in early semantic and syntax development” in D. Aaronson and R.W. Rieber (eds) Developmental psycholinguistics and communication disorders. Annals of the New York Academy of Sciences, 132–139. 195. Nerlich, B. and D. Clarke (1999) “Synecdoche as a Cognitive and Communicative Strategy” in A. Blank and P. Koch (eds) Historical Semantics and Cognition. Berlin: Mouton de Gruyter, 197–214. 196. Nerlich, B. and D. Clarke (2001) “Ambiguities I Live by: Toward a Pragmatics of Polysemy” in Journal of Pragmatics, 1–20. 197. Nerlich, B. and D. Clarke (2002) “Elements for an Integral Theory of Semantic Change and Semantic Development” in Studia Anglica Resoviensia 2, Wydawnictwo UR, 71–84. 198. Nice, M. (1925) “Length of Sentences as a Criterion of Child’s Progress in Speech” in Journal of Educational Psychology 16, 370–379. 199. Park, T. (1970a) “Language Acquisition of Korean Child”. Unpublished paper. University of Bern, Switzerland, Psychological Institute. 200. Park, T. (1970b) “The Acquisition of German Syntax”. Unpublished paper. University of Bern, Switzerland, Psychological Institute. 201. Peirce, Ch.S. (1965) Collected Papers. Cambridge, MA: Harvard University Press. 202. Peters, A.M. (1977) “Language Learning Strategies: Does the Whole Equal the Sum of the Parts?” in Language 53, 560–573. 203. Piaget, J. (1952) The Child’s Conception of Number. New York: Norton. 204. Piaget, J. (1954) The Construction of Reality in the Child. New York: Basic Books. 205. Piaget, J. (1963) “Explanation in Psychology and Psychophysiological Parallelism” in J. Piaget, P. Fraisse and M. Reuchlin (eds) Experimental Psychology: Its Scope and Method. New York: Basic Books, 153–192 206. Piaget, J. (1970) Genetic Epistemology. New York: Harper and Row. 207. Piaget, J. and B. Ihelder (1967) The Child’s Conception of Space. New York: Norton. 208. Piatelli-Palmarini, M. (1989) “Evolution, Selection and Cognition: from ‘Learning’ to Parameter Setting in Biology and in the Study of Language” in Cognition 31, 1–44. 209. Pine, J. and E. Lieven (1993) “Reanalyzing Rote-learned Phrases: Individual Differences in the Transition to Multi-word Speech” in Journal of Child Language 20, 551–571. 210. Pine, J., E. Lieven and C. Rowland (1998) “Comparing Different Models of the Development of the English Verb Category”in Linguistics 36, 4–40. 211. Pinker, S. (1984) “Visual Cognition: an Introduction” in Cognition 18, 1–64. 212. Pinker, S. (1994) The Language Instinct. London: Penguin Books. 213. Plunkett, K. and V. Marchman (1991) “U-shaped Learning and Frequency Effects in a Multilayered Perception: Implications for Child Language Acquisition” in Cognition 38, 73–193. 214. Poeppel, D. and Wexler, K. (1993) “The Full Competence Hypothesis of Clause Structure” in Language 69, 1–33. 215. Preyer, W. (1882) Die Seele des Kindes, Leipzig, published in English as The Mind of the Child, transl. H.W. Brown, New York: Appleton. 216. Radden, G. And K. Panther (2004) “Introduction: Reflections on Motivation” in Radden Günter and Klaus-Uwe Panther (eds) Studies in Linguistic Motivation (Cognitive Linguistics Research 28). Mouton de Gruyter: Berlin and New York, 3–46. 217. Radford, A. (1990) Syntactic Theory and the Acquisition of English Syntax. Oxford: Basil Blackwell. 218. Radford, A. (1997) “A Sociolinguistic Perspective on L2 Communication Strategies” in G. Kasper and E. Kellerman (eds) Communication Strategies: Psycholinguistic and Sociolinguistic Perspectives. London: Longman, 279–303.
211
219. Retheford, K., B. Schwartz and R. Chapman (1981) “Semantic Roles and Residual Grammatical Categories in Mother and Child Speech: Who Tunes in to Whom?” in Journal of Child Language 8, 583–608. 220. Richards, B.J. (1990) Language Development and Idividual Differences: A Study of Auxiliary Verb Learning. Cambridge: Cambridge University Press. 221. Rodd, L. and M. Braine (1970) “Children’s Imitations of Syntactic Constructions as a Measure of Linguistic Competence” in Journal of Verbal Learning and Verbal Behaviour 10, 430–443. 222. Rosch E. and B.B. Lloyd (eds) (1978) Cognition and Categorisation. Hillsdale, NJ: Lawrence Erlbaum Associates. 223. Rosch, E. (1977) “Classification of Real-world Objects: Origins and Representations in Cognition” in P. Johnson-Laird and P.Watson (eds) Thinking: Readings in Cognitive Science. Cambridge: Cambridge University Press, 501–519. 224. Rosch, E. (1978) “Principles of Categorization” in E. Rosch and B.B. Lloyd (eds) Cognition and Categorisation. Hillsdale, N.J.: Lawrence Erlbaum Associates, 27–48. 225. Rosch, E., C. Simpson and R.S. Miller (1976) “Structural Bases of Typicality Effects” in Journal of Experimental Psychology: Human Perception and Performance 2, 491–502. 226. Rozwadowski, J. (1903) “Semazyologia czyli nauka o rozwoju znaczeń wyrazów. Jej stan obecny, zasady i zadania” in Eos 17–111. 227. Rudzka-Ostyn, B. (ed.) (1988) Topics in Cognitive Linguistics. Amsterdam and Philadelphia: Benjamins. 228. Rydin, I. (1971) “A Swedish Child in the Beginning of Syntactic Development and some Cross-linguistic Comparisons”. Unpublished paper. Harvard University, Cambridge, Mass. 229. Sabourard, O. (1995) Le langage et ses maux. Paris: Editions Odile Jacob. 230. Sanches, M. (1978) Features in the Acquisition of Japanese Grammar. Unpublished doctoral book. Stanford, CA: Stanford University. 231. Schaerlaekens, A. (1973) The Two-word Sentence in Child Language Development. The Hague: Mouton. 232. Schiefelbusch, R.L. and L. Loyd (eds) (1974) Language Perspectives: Acquistion, Retardation and Intervention. Baltimore: University Park Press. 233. Schlesinger, I.M. (1971) “The Production of Utterances and Language Acquisition“ in D. Slobin (ed.), 61–101. 234. Schlesinger, I.M. (1974) “Relational Concepts Underlying Language” in R.L. Schiefelbusch and L.Loyd (eds), 129–151. 235. Schlesinger, I.M. (1982) Steps to language: Toward a Theory of Native Language Acquistion. Hillsdale, NJ: Lawrence Erlbaum Associates, 121–178. 236. Scholes, R. (1969) “The Role of Grammaticality in the Imitation of Woed Strings by Children and Adults” in Journal of Verbal Learning and Verbal Behaviour 8, 225–228. 237. Scholes, R. (1970) “On Functors and Contentives in Children’s Imitations of Word Strings” in Journal of Verbal Learning and Verbal Behaviour 9, 167–170. 238. Scollon, R. (1976) Conversations with a One-year-old: A Case Study of the Developmental Foundation of Syntax. Honolulu: The University of Hawaii Press. 239. Selinker, L. (1992) Rediscovering Interlanguage. London: Longman. 240. Shipley E., C. Smith and L. Gleitman (1969) “A Study in the Acquistion of Language: From Responses to Commands” in Language 45, 322–342. 241. Sinclair-de-Zwart, H (1971) “Sensorimotor Action Patterns as a Condition for the Acquistion of Syntax.” in R. Huxley and E. Ingram (eds) Language Acquisition: Models and Methods. New York: Academic Press, 103–114.
212
242. Singer, W. (2000) “Response Synchronization: A Universal Coding Strategy for the Definition of Relations” in M.S. Gazzaniga (ed.) The New Cognitive Neurosciences, Cambridge, MA: the MIT Press, 143–182. 243. Sinha, C., A. Thornseng, M. Hayashi and K.Plunkett (1994) “Comparative Spatial Semantics and and Language Acquisition: Evidence from Danish, English and Japanese” in Journal of Semantics 11, 253–287. 244. Skinner, B. (1957) Verbal Behaviour. New York: Appleton-Century-Crofts. 245. Slobin, D. (1970) “Universals of Grammatical Development in Children” in G.B. Flores d’Arcais and W. Levelt (eds) Advances in Psycholinguistics. Amsterdam: North Holland, 175–208. 246. Slobin, D. (ed.) (1971) The Ontogenesis of Grammar. New York: Academic Press. 247. Slobin, D. (1971) “On the Learning of Morphological Rules” in D. Slobin (ed.). 248. Slobin, D. (1973) “Cognitive Prerequisites for the Acquisition of Grammar” in D. Slobin and Ch. Ferguson, (eds) Studies of Child Language Development. New York: Holt, Rienhart and Winston, 173–208. 249. Slobin, D. (1979) Psycholinguistics. San Francisco: Scott Foresman and Co. 250. Slobin, D. (ed.) (1985) The Crosslinguistic Study of Language Acquisition. Hillsdale, NJ: Lawrence Erlbaum Associates. 251. Slobin, D. and Ferguson, Ch. (eds) (1973) Studies of Child Language Development. New York: Holt, Rienhart and Winston. 252. Smoczyńska, M. (1985) “The Acquisition of Polish” in D. Slobin (ed.), 595–686. 253. Southgate, V. and Meints, K. (2000) “Typicality, Labeling and Perceptual Contrast in Infants’ Category Membership Decisions: A Synergistic Effect” in Cognitive Linguistics 11, 4–16. 254. Staff, L. (2005) Poezje. Kraków: Zielona Sowa. 255. Steinberg, D. (1993) An Introduction to Psycholinguistics. London: Longman. 256. Stern, C. (1983) “The Prosody of Maternal Speech: Infant Age and Context Related Changes” in Journal of Child Language 10, 1–15. 257. Stern, C. (1907) Die Kindersprache: Eine Psychologische und Sprachteoretische Untersuchung. Leipzig: Barth. 258. Szwedek, A. (1976) Word Order, Sentence Stress and Reference in English and in Polish. Edmonton (Canada): Linguistic Research. 259. Szymanek, B. (1998) Introduction to Morphological Analysis. Warszawa: PWN. 260. Tabakowska, E. (1995) Gramatyka i obrazowanie. Wprowadzenie do językoznawstwa kognitywnego. Kraków: PAN. 261. Tabakowska, E. (2004) Kognitywizm po polsku: wczoraj i dziś. Kraków: Universitas. 262. Tabakowska, E. (ed.) (2001) Kognitywne podstawy języka i językoznawstwa. Kraków: Universitas. 263. Taine, H. (1877) “Acquisition of Language by Children” in Mind 2, 252–259. 264. Talmy, L. (1988) “The Relation of Grammar to Cognition” in B. Rudzka-Ostyn (ed.), 165– 205. 265. Taylor, J.R. (1989) Linguistic Categorization. Prototypes in Linguistic Theory. Oxford: Clarendon Press. 266. Taylor, J.R. (2002) Cognitive Grammar. Oxford: Oxford University Press. 267. Thorndike, E. (1932) The Fundamentals of Learning. New York: Columbia Teachers College. 268. Tiedemann, D. (1787) Beobachtungen über die Entwicklung der Seelenfähigkeit bei Kindern. Altenburg.
213
269. Tomasello, M (1995) “Language is not an Instinct” in Cognitive Development 10, 131–156. 270. Tomasello, M. (1992) First Verbs: A Case Study in Early Grammatical Development. Cambridge: Cambridge University Press. 271. Tomasello, M. (1998a) “The Return of Constructions. Review of Adele Goldberg ‘Constructions’. A Construction Grammar Approach to Argument Structure” in Journal of Child Language 25, 431–442. 272. Tomasello, M. (1998b) “Reference: Intending that Others Jointly Attend” in Pragmatics and Cognition 6, 219–234. 273. Tomasello, M. (1999) The Cultural Origins of Human Cognition. Harvard: Harvard University Press. 274. Tomasello, M. (2000) “Do Young Children Have Adult Syntactic Competence?” in Cognition 74, 209–253. 275. Tomasello, M. (2003) Constructing a Language: A Usage-Based Theory of Child Language Acquistion. Harvard: Harvard University Press. 276. Van Valin, R. (1991) “Functionalist Linguistic Theory and Language Acquisition” in First Language 11, 7–40. 277. Watson, J. (1924) Behaviourism. New York: Norton. 278. Weissenborn, J. (1990) “Functional Categories and Verb Movement: The Acquistion of German Syntax Reconsidered” in M Rothweiler (ed.) Spracherwerb und Grammatik, Opladen: Westdeutcher Verlag, 190–224. 279. Wells, G. (1985) Language Development in the Preschool Years. Cambridge: Cambridge University Press. 280. Westfal, S. (1956) A Study in Polish Morphology: The Genitive Singular Masculine. The Hague: Mouton. 281. Wexler, K. (1994) “Optional Infinitives, Head Movement, and the Economy of Derivation” in D. Lightfoot and N.Hornstein (eds) Verb Movement. Cambridge: Cambridge University Press, 305–350. 282. Wexler, K. and P. Culicover (1980) Formal Principles of Language Acquisition. Cambridge, MA: the MIT Press. 283. Wieman, L. (1976) “Stress Patterns of Early Child Language” in Journal of Child Language 3, 283–286. 284. Wierzbicka, A. (1988). The Semantics of Grammar. Amsterdam: John Benjamins. 285. Wundt, W. (1900) Volkerpsychologie; Eine Untersuchung der Entwicklungsgesteze von Sprache. Leipzig: Kröner. 286. Wurzel, W. (1994) “Natural Morphology” in R.E. Ascher and J.M.Y. Simpson (eds) The Encyclopaedia of Language and Linguistics. Oxford: Pergamon Press, 2590–8.
Reference books 287. Dictionary of Language Teaching and Applied Linguistics (1999). London: Longman. 288. Hutchinson Encyclopaedia (2000) Oxford: Helicon Publishing LTD. 289. Malmkjær, K. (ed.) (2004) The Linguistics Encyclopedia. London and New York: Routledge. 290. Polański, K. (ed.) (1971) Encyklopedia językoznawstwa ogólego. Wrocław: Ossolineum.
214
