The dimension of emotion in words is also starting to be understood ... and the affect of the writer, characters in a story may use specific lexical terms to denote a certain ... will not be accepted as full partners without displaying some humor capabilities of their own. ... For example, the persuasive effect of humor and emotions ...
Moving Creative Words Oliviero Stock, Carlo Strapparava, and Alessandro Valitutti Istituto per la Ricerca Scientifica e Tecnologica, FBK-irst, I-38050, Povo, Trento, ITALY {stock, strappa, alvalitu}@itc.it
Abstract. Among forms of creative language, verbal humor has received some attention in the computational milieu. Some aspects of irony and wordplay could be experimented in automated systems. For instance we developed a system that makes fun of existing acronyms, based mainly on lexical reasoning. The dimension of emotion in words is also starting to be understood among computational linguists. The challenge of electronic advertisements offers in particular a great opportunity for getting now deeper into creative language expression and emotion. An advertising message induces in the recipient a positive or negative attitude toward the object to advertise. A prototype we have developed for advertising professionals has two steps: (i) the creative variation of familiar expressions, taking into account the affective content of the produced text, (ii) the automatic animation (semantically consistent with the affective text content) of the resulting expression, using kinetic typography techniques. Validation prospects are also challenging and will be briefly discussed.
1
Introduction
In recent times the landscape of natural language processing has been enriched with elements of emotion-related processing. A text often reflects the opinions and the affect of the writer, characters in a story may use specific lexical terms to denote a certain emotional state, a dialogue is strongly affected by the evolution of the affective states of the participants. One puzzling topic at the border of affective communication is humor. Humor has been studied since ancient times and in the Twentieth Century various theories have been proposed in fields such as philosophy, linguistics, and psychology. Yet, a deep understanding of the mechanisms is beyond the state of the art. We believe the computational approach can contribute something here, as it has happened with other areas of artificial intelligence. We also believe that verbal humor touches on aspects of aesthetics and also from that point of view a realization of some of the linguistic creative processes may help us understand communication in a broad sense. If we look at things from a different, applied perspective we can well say that humor is a need in our relation with our fellow humans. As computers will be able to yield autonomous contributions to communication, they will not be accepted as full partners without displaying some humor capabilities of their own.
In concrete terms computational humor has the potential to change computers into extraordinarily creative and motivational tools. Computer-human interaction needs to evolve beyond usability and productivity. There is a wide perception in the field that the future is in themes such as entertainment, fun, emotions, aesthetic pleasure, motivation, attention, engagement and so on. Humor is an essential element in communication: it is strictly related to the themes mentioned above. While it is generally considered merely a way to induce amusement, humor provides an important way to influence the mental state of people to improve their activity. Even though humor is a very complex capability to reproduce, it is realistic to model some types of humor production and to aim at implementing this capability in computational systems. Let us now review a few elements that make humor so important from a cognitive point of view. Humor and emotions. Humor is a powerful generator of emotions. As such, it has an impact on people’s psychological state, directs their attention [1], influences the processes of memorization [2] and of decision-making [3], and creates desires. Actually, emotions are an extraordinary instrument for motivation and persuasion because those who are capable of transmitting and evoking them have the power to influence other people’s opinions and behaviour. Humor, therefore, allows for conscious and constructive use of the affective states generated by it. Affective induction through verbal language is particularly interesting; and humor is one of the most effective ways of achieving it. Purposeful use of humorous techniques enables us to induce positive emotions and mood and to exploit their cognitive and behavioural effects. For example, the persuasive effect of humor and emotions is well known and widely employed in advertising. Advertisements have to be both short and meaningful, to be able to convey information and emotions at the same time. Humor and beliefs. Humor acts not only upon emotions, but also on human beliefs. A joke plays on the beliefs and expectations of the hearer. By infringing on them, it causes surprise and then hilarity. Jesting with beliefs and opinions, humor induces irony and accustoms people not to take themselves too seriously. Sometimes simple wit can sweep away a negative outlook that places limits on people desires and abilities. Wit can help people overcome self-concern and pessimism that often prevents them from pursuing more ambitious goals and objectives. Humor and creativity. Humor encourages creativity as well. The change of perspective caused by humorous situations induces new ways of interpreting the same event. By stripping away clich´es and commonplaces, and stressing their inconsistency, people become more open to new ideas and points of view. Creativity redraws the space of possibilities and delivers unexpected solutions to problems. Actually, creative stimuli constitute one of the most effective impulses for human activity. Machines equipped with humorous capabilities will be able
to play an active role in inducing users’ emotions and beliefs, and in providing motivational support.
2
Background
While humor is relatively well studied in scientific fields such as linguistics [4] and psychology [5, 6], to date there is only a limited number of research contributions made toward the construction of computational humor prototypes. A good review of the field can be found in [7]. Almost all the approaches try to deal with incongruity theory at various levels of refinement [8, 9, 4]. Incongruity theory focuses on the element of surprise. It states that humor is created out of a conflict between what is expected and what actually occurs in the joke. Underlying incongruity is one of the obvious features of a large part of humor phenomena: ambiguity or double meaning. One of the first attempts that deals with humor generation is the work described in [10], where a formal model of semantic and syntactic regularities was devised, underlying some types of puns (punning riddles). A punning riddle is a question-answer riddle that uses phonological ambiguity. The three main strategies used to create phonological ambiguity are syllable substitution, word substitution and metathesis. Syllable substitution is the strategy to confuse a syllable in a word with a similar or identical sounding word. An example of syllable substitution is shown in the following joke: “What do shortsighted ghosts wear? Spooktacles” [11]. Word substitution is the strategy to confuse an entire word with another similar- or identical-sounding word. An example of a joke with word substitution is : “How do you make gold soup? Put fourteen carrots in it” [11]. Metathesis is a strategy very different to syllable or word substitution. It uses reversal of sounds and words to suggest a similarity in meaning between two semantically distinct phrases. An example is “What is the difference between a torn flag and a postage stamp? One’s a tattered banner and the other’s a battered tanner.” [10]. Punning riddles based on these three strategies are all suitable for computer generation. Ritchie and Binsted focussed on the word substitution based punning riddles, as lists of homophones (i.e. phonetically identical words) are already available. The assumptions about the contents and the structure of the lexicon are as follows. The lexicon consists of a finite set of lexemes and of lexical relations. A lexeme is an abstract entity corresponding to the meaning of a word. If a word has two meanings, it has two corresponding lexemes. Every lexeme has a set of properties about the representation and the type of word. A lexical relation can be an explicit relation between two lexemes, like synonym or homophone, or a general inter-lexeme relation, applicable to more than one pair of lexemes. In order to describe a punning riddle, two sorts of symbolic description have to be used: schema and template. A schema stipulates a set of relations which must hold between the lexemes used to build a joke. A template indicates the information necessary to turn a schema and lexemes into a piece of text. It
contains fixed segments of text that are to be used and syntactic details of how lexemes have to be expressed. In [10], this model was then exploited to implement a system called JAPE, able to automatically generate amusing puns. In a recent work [12] automatic production of a funny and appropriate punchline at the end of short jokes is proposed. The authors present a model that describes the relationship between the connector (part of the set-up) and the disjunctor (the punchline). In particular they have implemented this model in a system which, given a joke set-up, can select the best disjunctor from a list of alternatives. Another humor-generation project was HAHAcronym [13], whose goal was to develop a system able to automatically generate humorous versions of existing acronyms, or to produce a new funny acronym constrained to be a valid vocabulary word, starting with concepts provided by the user. The humorous effect was achieved mainly on the basis of incongruity. We will provide examples of output of this system in Section 3. Humor recognition has received less attention. In [14] the application of text categorization techniques to humor recognition has been investigated. In particular the authors show that classification techniques are a viable approach for distinguishing between humorous and non-humorous text, through experiments performed on very large data sets. They restrict their investigation to the type of humor found in one-liners. A one-liner is a short sentence with comic effects and a peculiar linguistic structure: simple syntax, deliberate use of rhetoric devices (e.g. alliteration, rhyme), and frequent use of creative language constructions meant to attract the readers’ attention. In fact, while longer jokes can have a relatively complex narrative structure, a one-liner must produce the humorous effect “in one shot”, with very few words. The humor-recognition problem is formulated as a traditional classification task, feeding positive (humorous) and negative (non humorous) examples to some automatic classifiers. The humorous data set consisted of a corpus of 16,000 oneliners collected from the Web using an automatic bootstrapping process. The non-humorous data were selected such that it is structurally and stylistically similar to the one-liners. In particular, four different corpora were selected, each composed by 16,000 sentences: (1) Reuters news titles [15]; (2) proverbs; (3) sentences picked from the British National Corpus (BNC)[16]; and (4) commonsense statements from the Open Mind Common Sense (OMCS) corpus [17]. The features taken into account were both content-based features, usually considered in traditional text categorization tasks, and humor-specific stylistic features, such as alliteration, presence of antonymy and adult slang. The classification results were really encouraging. Regardless of the non-humorous data set playing the role of negative examples, the performance of the automatically learned humorrecognizer was always significantly better than apriori known baselines. Surprisingly, comparative experimental results showed that in fact it is more difficult to distinguish humor from regular text (e.g. BNC sentences) than from the other data sets.
Another related work is the study reported in [18], focussing on a very restricted type of wordplays, namely the “Knock-Knock” jokes. The goal of the study was to evaluate to what extent wordplay can be automatically identified in “Knock-Knock” jokes, and if such jokes can be reliably identified from other non-humorous texts. The algorithm is based on automatically extracted structural patterns and on heuristics heavily based on the peculiar structure of this particular type of jokes. While the wordplay recognition gave satisfactory results, the identification of jokes containing such wordplays turned out to be significantly more difficult. Worth mentioning is also a formalization, based on a cognitive approach (the belief-desire-intention model), distinguishing between real and fictional humor [19]. Finally [20] proposes a first attempt to recognize the humorous intent of short dialogs. According to the authors, computational recognition of humorous intent can be divided into two parts: recognition of a humorous text, and recognition of the intent to be humorous. The approach is based on detecting ambiguity both in the setup and in the punchline.
3
HAHAcronym
HAHAcronym was the first European project devoted to computational humor1 . The main goal of HAHAcronym was the realization of an acronym ironic reanalyzer and generator as a proof of concept in a focalized but non restricted context. In the first case the system makes fun of existing acronyms, in the second case, starting from concepts provided by the user, it produces new acronyms, constrained to be words of the given language. And, of course, they have to be funny. The realization of this system was proposed to the European Commission as a project that we would be able to develop in a short period of time (less than a year), that would be meaningful, well demonstrable, that could be evaluated along some pre-decided criteria, and that was conducive to a subsequent development in a direction of potential applicative interest. So for us it was essential that: 1. the work could have many components of a larger system, simplified for the current setting; 2. we could reuse and adapt existing relevant linguistic resources (e.g. WordNet Domains, assonance tools, parser, etc.); 3. some simple strategies for humor effects could be experimented. One of the purposes of the project was to show that using “standard” resources (with some extensions and modifications) and suitable linguistic theories of humor (i.e. developing specific algorithms that implement or elaborate theories), it is possible to implement a working prototype. 1
EU project IST-2000-30039 (partners: ITC-irst and University of Twente), part of the Future Emerging Technologies section of the Fifth European Framework Program.
3.1
Examples
Here below some examples of acronym re-analysis by HAHAcronym are reported. As far as semantic field opposition is concerned we have slightly tuned the system towards the domains Food, Religion and Sex. We report the original acronym, the re-analysis and some comments about the strategies followed by the system. ACM - Association for Computing Machinery → Association for Confusing Machinery FBI - Federal Bureau of Investigation → Fantastic Bureau of Intimidation The system keeps all the main heads and works on the adjectives and the PP head, preserving the rhyme and/or using the a-semantic dictionary. CRT - Cathodic Ray Tube → Catholic Ray Tube ESA - European Space Agency → Epicurean Space Agency PDA - Personal Digital Assistant → Penitential Demoniacal Assistant → Prenuptial Devotional Assistant MIT - Massachusetts Institute of Technology → Mythical Institute of Theology Some re-analyses are Religion oriented. Note the rhymes. As far as generation from scratch is concerned, a main concept and some attributes (in terms of Wordnet synsets) are given as input to the system. Here below we report some examples of acronym generation. Main concept: processor (in the sense of CPU); Attribute: fast OPEN - On-line Processor for Effervescent Net PIQUE - Processor for Immobile Quick Uncertain Experimentation TORPID - Traitorously Outstandingly Rusty Processor for Inadvertent Data processing UTMOST - Unsettled Transcendental Mainframe for Off-line Secured Tcp/ip We note that the system tries to keep all the expansions of the acronym coherent in the same semantic field of the main concept (Computer Science). At the same time, whenever possible, it exploits incongruity in the lexical choices.
4
Creative Messages and Optimal Innovation
Variating familiar expressions (proverbs, movie titles, famous citations, etc.) in an evocative way has been an effective technique in advertising for a long time [21]. A lot of efforts by professionals in the field goes into producing ever novel
catchy expressions with some element of humor. Indeed it is common of “creatives” to be recruited in pairs formed by a copywriter and an art director. They work in a creative partnership to conceive, develop and produce effective advertisement. While the copywriter is mostly responsible for the textual content of the creative product, the art director focalizes efforts on the graphical presentation of the message. Advertising messages tend to be quite short but, at the same time, rich of emotional meaning and persuasive power. We combined some computational functionalities for the semiautomatic production of creative advertising messages. In particular, we implemented a strategy for the creative variation of familiar expressions. This strategy is articulated in two steps. The first consists of the selection and creative variation of familiar or common sense expressions. The second step consists of the presentation of the headline through automated text animation, and it is based on the use of kinetic typography. An advertising message induces in the recipient a positive (or negative) attitude toward the subject to advertise, for example through the evocation of a appropriate emotion. Another mandatory characteristic of an advertisement is its memorizability. These two aspects of an ads increase the probability to induce some wanted behaviours, for example the purchase of some product, the choice of a specific brand, or the click on some specific web link. In the last case, it is crucial to make the recipient curious about the subject referred by the URL. The best way to realize in an ads both attitude induction and memorizability is the generation of surprise, generally based on creative constraints. In order to develop a strategy for surprise induction, we considered an interesting property of pleasurable creative communication that was named by Rachel Giora as the optimal innovation hypothesis ([22]). According to this assumption, when the novelty is in a complementary relation to salience (familiarity), it is “optimal” in the sense that it has an aesthetics value and “induce the most pleasing effect”. Therefore the simultaneous presence of novelty and familiarity makes the message potentially surprising, because this combination allows the recipient’s mind to oscillate between what is known and what is different from usual. For this reasons, an advertising message must be original but, at the same time, connected to what is familiar [21]. Familiarity causes expectations, while novelty violates them, and finally surprise arises. With “varied familiar expression” we indicate an expression (sentence or phrase) that is obtained as a linguistic change (e.g. substitution of a word, morphological or phonetic variation, etc.) of an expression recognized as familiar by recipients (e.g. selected by some collection of proverbs, famous movie titles, etc.). In this work we limited the variation to the word substitution. Moreover, a successful message should have a semantic connection with some concept of the target topic. At the same time, it has to be semantically related with some emotion of a prefixed valence (e.g. positive emotion as joy or negative emotion as fear) .
5 5.1
Resources Affective Semantic Similarity
All words can potentially convey affective meaning. Each of them, even those more apparently neutral, can evoke pleasant or painful experiences. While some words have emotional meaning with respect to the individual story, for many others the affective power is part of the collective imagination (e.g. words “mum”, “ghost”, “war” etc.). We are interested in this second group, because their affective meaning is part of common sense knowledge and can be detected in the linguistic usage. For this reason, we studied the use of words in textual productions, and in particular their co-occurrences with the words in which the affective meaning is explicit. As claimed by Ortony et al. [23], we have to distinguish between words directly referring to emotional states (e.g. “fear”, “cheerful”) and those having only an indirect reference that depends on the context (e.g. words that indicate possible emotional causes as “killer” or emotional responses as “cry”). We call the former direct affective words and the latter indirect affective words [24]. In order to manage affective lexical meaning, we (i) organized the direct affective words and synsets inside WordNet-Affect, an affective lexical resource based on an extension of WordNet, and (ii) implemented a selection function (named affective weight) based on a semantic similarity mechanism automatically acquired in an unsupervised way from a large corpus of texts (100 millions of words), in order to individuate the indirect affective lexicon. Applied to a concept (e.g. a WordNet synset) and an emotional category, this function returns a value representing the semantic affinity with that emotion. In this way it is possible to assign a value to the concept with respect to each emotional category, and eventually select the emotion with the highest value. Applied to a set of concepts that are semantically similar, this function selects subsets characterized by some given affective constraints (e.g. referring to a particular emotional category or valence). As we will see, we are able to focus selectively on positive, negative, ambiguous or neutral types of emotions. For example, given “difficulty” as input term, the system suggests as related emotions: identification, negative-concern, ambiguous-expectation, apathy. Moreover, given an input word (e.g. “university”) and the indication of an emotional valence (e.g. positive), the system suggests a set of related words through some positive emotional category (e.g. “professor” “scholarship” “achievement”) found through the emotions enthusiasm, sympathy, devotion, encouragement. This fine-grained affective lexicon selection can open up new possibilities in many applications that exploit verbal communication of emotions. For example, [25] exploited the semantic connection between a generic word and an emotion for the generation of affective evaluative predicates and sentences.
5.2
Database of Familiar Expressions
The base for the strategy of “familiar expression variation” is the availability of a set of expressions that are recognized as familiar by English speakers. We considered three types of familiar expressions: proverbs, movie titles, clich´es. We collected 1836 familiar expressions from the Web, organized in three types: common use proverbs (628), famous movie titles (290), and clichg´es (918). Proverbs were retrieved in some of many web sites in which they are grouped (e.g. http://www.francesfarmersrevenge.com/stuff/proverbs.htm or www.manythings.org /proverbs). We considered only proverbs of common use. In a similar way we collected clich´es, that are sentences whose overuse often makes them humorous (e.g. home sweet home, I am playing my own game). Finally, movie titles were selected from the Internet Movie Database (www.imdb.com). In particular, we considered the list of the best movies in allo sorts of categories based on votes from users. The list of familiar expressions is composed mostly of sentences (in particular, proverbs and clich´es), but part of them are phrases (in particular, movie title list includes a significant number of noun phrases) 5.3
Assonance Tool
To cope with this aspect we got and reorganized the CMU pronouncing dictionary (http://www.speech.cs.cmu.edu/cgi-bin/cmudict) with a suitable indexing. The CMU Pronouncing Dictionary is a machine-readable pronunciation dictionary for North American English that contains over 125,000 words and their transcriptions. Its format is particularly useful for speech recognition and synthesis, as it has mappings from words to their pronunciations in the given phoneme set. The current phoneme set contains 39 phonemes; vowels may carry lexical stress. 5.4
Kinetic Typography Scripting Language
Kinetic typography is the technology of text animation, i.e. text that uses movement or other changes over time. The advantage of kinetic typography consists in a further communicative dimension, combining verbal and visual communication, and providing opportunities to enrich the expressiveness of static texts. According to [26], kinetic typography can be used for three different communicative goals: capturing and directing attention of recipients, creating characters, and expressing emotions. A possible way of animating a text is mimicking the typical movement of humans when they express the content of the text (e.g. “Hi” with a jumping motion mimics exaggerated body motion of humans when they are really glad). We have realized a development environment for the creation and visualization of text animations based on Kinetic Typography Engine (KTE), a Java package developed at the Design School of Carnegie Mellon University [26].Our model for the animation representation is a bit simpler than the KTE model. The
central assumption consists of the representation of the animation as a composition of elementary animations (e.g. linear, sinusoidal or exponential variation). In particular, we consider only one operator for the identification of elementary animations (k-base) and three composition operators: kinetic addition (k-add), kinetic concatenation (k-join), and kinetic loop (k-loop).
6
Algorithm
In this section, we describe the algorithm developed to perform the creative variation of an existing familiar expression. 1. Insertion of an input concept. The first step of the procedure consists of the insertion of an input concept. This is represented by one or more words, a set of synonyms, or a WordNet synset. In the latter case, it is individuated through a word, the part of speech (noun, adjective, verb, or adverb), and the sense number, and it corresponds to a set of synonyms. Using the pseudodocument representation technique described above, the input concept is represented as a vector in the LSA vectorial space. For example, say that a cruise vacation agency seeks to produce a catchy message on the topics “vacation” and “beach”. 2. Generation of the target-list. A list (named target-list) including terms that are semantically connected (in the LSA space) with the input concept(s) is generated. This target list represents a semantic domain that includes the input concept(s).For example, given the vector representing “vacation”, “beach”, the LSA returns a list “sea”, “hotel”, “bay”, “excursion”, etc. 3. Association of assonant words. For each word of the target-list one or more possible assonant words are associated. Then a list of word pairs (named variation-pairs) is created. The list of variation-pairs is filtered according to some constraints. The first one is syntactic (elements of each pair must have the same part of speech). The second one is semantic (i.e. the second element of each pair must not be included in the target-list), and its function is to realize a semantic opposition between the elements of a variation pair. Finally, to each variation pair an emotion-label (representing the emotional category most similar to the substituting word) is provided with the corresponding affective weight. Some possible assonant pairs for the example above are: (bay, day), (bay, hay), (hotel, farewell), etc. 4. Creative variation of familiar expressions. In this step, the algorithm gets in input a set of familiar expressions (in particular, proverbs and movie titles) and, for each of them, generates all possible variations. The list of variated expressions is ordered according to the global affective weight. Following the example, a resulting ad is Tomorrow is Another Bay as a variation of the familiar expression Tomorrow is Another Day. Note that for moment the final choice among the best resulting expressions proposed by the system is left to human selection.
At this point, the variated expression is animated with kinetic typography. In particular, words are animated according to the underlying emotion to emphasize the affective connotation.
7
Examples
In this section we show some examples of creative variations. Starting from an input concept (e.g. disease) we can obtain, using the semantic similarity, a list of related terms (Table 1).
Name POS Similarity to the input symptom noun 0.971 therapy noun 0.969 metabolism noun 0.933 analgesic noun 0.899 suture noun 0.851 thoracic adjective 0.782 extraction noun 0.623 Table 1. Input word: “disease”
Name fear joy anger sadness disease 0.357 0.201 0.135 0.679 symptom 0.423 0.293 0.164 0.685 therapy 0.374 0.315 0.170 0.691 metabolism 0.372 0.258 0.082 0.552 analgesic 0.280 0.241 0.173 0.526 suture 0.237 0.299 0.227 0.490 thoracic 0.157 0.135 0.134 0.448 extraction 0.126 0.245 0.177 0.366 Table 2. Affective weight
Using the affective weight function, it is possible to check for their affective characterization (in Table 2 only four emotions are displayed), selecting those affectively coherent with the input term. Subsequently, the system searches for assonant words (Table 3) and checks for affective opposition with the original words (Table 4). At this point, the system retrieves familiar expressions that include the word to be substituted.
Name suture thoracic extraction
Assonant Words future Jurassic abstraction, attraction, contraction, diffraction, distraction, inaction, reaction, retraction, subtraction, transaction Table 3. Phonetic associations
Name fear joy anger sadness suture 0.237 0.299 0.227 0.490 future 0.467 0.571 0.417 0.462 Table 4. Affective difference
Input Words Varied Expression Word Substitution vacation, beach Tomorrow is another bay day → bay disease Back to the Suture future → suture Thoracic Park jurassic → thoracic Fatal Extraction attraction → extraction crash Saturday Fright Fever night → fright fashion Jurassic Dark park → dark Table 5. More Examples
Table 5 shows the final word substitution in several examples. The system can then automatically animate the resulting expression emphasizing the novel affective connotation through kynetic typography techniques as shown in [27].
8
Humor and Neuroimaging
Deep evaluation of achieved results is not an easy task. Normally it is performed with user’s direct feedback. Recent advances in cognitive neuroscience are worth examining as a potential new approach. In particular, there are a number of experiments of functional neuroimaging aimed at individuating neural correlates of humour comprehension and appreciation. These results were compared to studies on patients with brain lesions, leading in some cases to different outcomes, but in general the cognitive model was validated (for a complete review, see [28],). Generally the framework within which neuroimaging studies are interpreted is the Incongruity-Resolution Theory of Humour [29]. It is based on a two-stage model of humour comprehension. The first stage is the detection of an incongruity in some joke or pun. Incongruity is perceived when some expectation is disconfirmed and surprise arises. The second stage is the reinterpretation of the situation expressed in the text in a way that is congruous and funny.
Illustrative of the neuroimaging approach to humour are experiments by Mobbs et al. [30] and Bartolo et al. [31], based on event-related functional MRI (efMRI) study of humour comprehension. Both studies aimed at measuring hemodynamic increases in regions associated with cartoons considered to be funny. The results are coherent with previous analog experiments, and allow us to identity different clusters of brain areas with a significant BOLD signal, corresponding to the cognitive-affective components of humour comprehension: humour detection (including incongruity detection and incongruity resolution), motor response and affective response. The most important feature of humour appreciation is reward, the amusement that follows the humorous stimulus. At the moment there are not results that conclusively demonstrate the subcortical correlates of reward, but there are a number of fMRI studies on different rewarding tasks (for review, see Schultz [32]). Functional neuroimaging of humour appreciation could be useful for the evaluation of computational humour systems. The possibility of integrating information coming from subjects reports and direct neural functional activity is certainly appealing.
9
Conclusion
In this paper, we have presented some recent developments in automatic verbal humor production. We have described a prototype that produces creative variation of familiar expressions, exploiting state-of-the-art natural language processing techniques, and animates them according to the affective content. The creative textual variations rely on semantic and affective similarity, while animation makes use of a kinetic typography dynamic scripting language. The multimodal dynamic result is supposed to have a stronger effect. Evaluation is still preliminary and it may be worth looking into novel methodologies for appreciating the effects.
Acknowledgments This work was developed in the context of HUMAINE Network of Excellence and partially sponsored by MUR FIRB-project number RBIN045PXH.
References 1. Kitayama, S., Niedenthal, P.: The Heart’s Eye: Emotional influences in Perception and Attention. Academic Press (1994) 2. Kahneman, D.: Attention and Effort. Englewood Cliffs: Prentice-Hall (1973) 3. Isen, A.: Positive affect and decision making. In Lewis, M., Haviland, J., eds.: Handbook of Emotion. Guilford, New York (1993) 4. Attardo, S.: Linguistic Theories of Humor. Mouton de Gruyter, Berlin (1994)
5. Freud, S.: Der Witz und Seine Beziehung zum Unbewussten. Deutike, Leipzig and Vienna (1905) 6. Ruch, W.: Computers with a personality? lessons to be learned from studies of the psychology of humor. In: [33]. (2002) 7. Ritchie, G.: Current directions in computational humour. Artificial Intelligence Review 16(2) (2001) 119–135 8. Koestler, A.: The act of creation. Hutchinson, London (1964) 9. Raskin, V.: Semantic Mechanisms of Humor. Dordrecht/Boston/Lancaster (1985) 10. Binsted, K., Ritchie, G.: Computational rules for punning riddles. Humor 10(1) (1997) 11. Webb, K., ed.: The Crack-a-Joke Book. Puffin, London (1978) 12. Stark, J., Binsted, K., Bergen, B.: Disjunctor selection for one-line jokes. In: Proc. of Intelligent Technologies for Interactive Entertainment (INTETAIN ’05), Madonna di Campiglio, Italy (December 2005) 13. Stock, O., Strapparava, C.: Getting serious about the development of computational humour. In: Proceedings of the 18th International Joint Conference on Artificial Intelligence (IJCAI-03), Acapulco, Mexico (August 2003) 14. Mihalcea, R., Strapparava, C.: Laughter abounds in the mouths of computers: Investigations in automatic humor recognition. In: Proc. of Intelligent Technologies for Interactive Entertainment (INTETAIN ’05), Madonna di Campiglio, Italy (December 2005) 15. Lewis, D., Yang, Y., Rose, T., Li, F.: RCV1: A new benchmark collection for text categorization research. The Journal of Machine Learning Research 5 (December 2004) 361–397 16. BNC-Consortium: British national corpus http://www.hcu.ox.ac.uk/bnc/. (2000) 17. Singh, P.: The public acquisition of commonsense knowledge. In: Proceedings of AAAI Spring Symposium: Acquiring (and Using) Linguistic (and World) Knowledge for Information Access., Palo Alto, CA (2002) 18. Taylor, J., Mazlack, L.: Computationally recognizing wordplay in jokes. In: Proceedings of CogSci 2004, Chicago (August 2004) 19. Mele, F.: Real and fictional ridicule. In: [33]. (2002) 20. Taylor, J., Mazlack, L.: Toward computational recognition of humorous intent. In: Proc. of the 27th Annual Conference of the Cognitive Science Society (COGSCI 05), Stresa, Italy (July 2005) 21. Pricken, M.: Creative Advertising. Thames & Hudson (2002) 22. Giora, R.: On Our Mind: Salience, Context and Figurative Language. Oxford University Press, New York (2003) 23. Ortony, A., Clore, G.L., Foss, M.A.: The psychological foundations of the affective lexicon. Journal of Personality and Social Psychology 53 (1987) 751–766. 24. Strapparava, C., Valitutti, A., Stock, O.: The affective weight of lexicon. In: Proceedings of the Fifth International Conference on Language Resources and Evaluation (LREC 2006), Genoa, Italy (May 2006) 25. Valitutti, A., Strapparava, C., Stock, O.: Lexical resources and semantic similarity for affective evaluative expressions generation. In: Proc. of the First International Conference on Affective Computing & Intelligent Interaction (ACII 2005), Beijing, China (October 2005) 26. Lee, J., Forlizzi, J., Hudson, S.: The kinetic typography engine: An extensible system for animating expressive text. In: Proc. of ACM UIST 2002 Conference. (2002)
27. Strapparava, C., Valitutti, A., Stock, O.: Dances with words. In: Proceedings of the 20th International Joint Conference on Artificial Intelligence (IJCAI-07), Hyderabad, India (January 2007) 28. Wild, B., Rodden, F.A., Grodd, W., Ruch, W.: Neural correlates of laughter and humour. Brain (126) (2003) 2121–2138 29. Suls, J.M.: A two-stage model for the appreciation of jokes and cartoons. In Goldstein, P.E., McGhee, J.H., eds.: The psychology of humour. Theoretical perspectives and empirical issue. Academic Press, New York (1972) 81–100 30. Mobbs, D., Greicius, M.D., Abdel-Azim, E., Menon, V., Reiss, A.L.: Humor modulates the mesolimbic reward centers. Neuron 40 (2003) 1041–1048 31. Bartolo, A., Benuzzi, F., Nocetti, L., Baraldi, P., Nichelli, P.: Humor comprehension and appreciation: An fmri study. Journal of Cognitive Neuroscience 18(11) (2006) 1789–1798 32. Schultz, W.: Getting formal with dopamine and reward. Neuron (36) (2002) 241– 263 33. Stock, O., Strapparava, C., Nijholt, A., eds.: Proceedings of the The April Fools Day Workshop on Computational Humour (TWLT20), Trento (2002)
