AUTHOR’S COPY | AUTORENEXEMPLAR
Relevance theory and unintended transmission of information José María Gil
We see it as a major challenge for any account of human communication to give a precise description and explanation of its vaguer effects. Distinguishing meaning from communication, accepting that something can be communicated without being strictly speaking meant by the communicator or the communicator’s behaviour, is a first essential step. —Dan Sperber and Deirdre Wilson (1995: 57)
Abstract In spite of the epigraph, the Theory of Relevance has overestimated the importance of intention in the study of verbal interaction. Therefore, this prototypical and current representative of Cognitive-Philosophical Pragmatics has inconveniently narrowed the study of language use. On the other hand, Neurocognitive Linguistics enables us to consider verbal interaction in wider and more realistic terms.
1. Introduction There are several trends in pragmatics that aim to account for non-communicative forms of language and many other aspects of language use (Bierwisch 1980; Kasher 1986, 1991; Nida 1990; Németh T. 2004, 2008). However, different trends of Cognitive-Philosophical Pragmatics, one of the most important and prestigious traditions within the study of language use, has not considered deeply some data that reveal important information about the structure of the cognitive system that underlies linguistic abilities. For example, after hearing Intercultural Pragmatics 8-1 (2011), 1–40 DOI 10.1515/IPRG.2011.001
1612-295X/11/0008-0001 © Walter de Gruyter
AUTHOR’S COPY | AUTORENEXEMPLAR
AUTHOR’S COPY | AUTORENEXEMPLAR
2 José María Gil that the sixth-grade teacher was going to be absent for a three-month period, one of the fathers attending a meeting said: (1) We need a prostitute teacher. He said “prostitute” instead of “substitute,” probably conditioned by the rhyme and other associations. This “Freudian slip” was produced spontaneously: The evidence for that is that the producer of (1) was surprised both when his hearers laughed and when the slip was pointed out. According to the traditional pragmatic analysis inaugurated by Grice in his famous work on conversational implicatures (1967), such an example is not an instance of verbal communication simply because the speaker did not have the intention to mean anything about any prostitute. The producer of (1) probably transmitted some information about his feelings or thoughts, but this transmission of information is linguistically irrelevant for orthodox Gricean pragmatics: It could be compared with the information that is unintentionally trans mitted by the trembling of the voice. The Theory of Relevance is one of the most important pragmatic models developed on the basis of Gricean insights on conversation. In this context, the slip of the tongue produced in (1) is not ostensive; consequently, it is not the result of a communicative intention. In Sperber-Wilson’s words, the producer of (1) did not want to make mutually manifest to audience and himself that he had the intention to inform the necessity of a prostitute. It must be admitted that, although slips of the tongue or unintended puns are not instances of ostensive behavior, they always occur during ostensive behavior. In fact, slips of the tongue and unintended puns coexist with ostensive-inferential communication. Within this context, it should be emphasized that unintended transmission of information (promoted for example by a slip of the tongue or an unintended pun) does not exclude ostensive-inferential communication, and vice versa. Probably, the two processes are always together. It has been widely recognized that the Theory of Relevance accounts for fundamental aspects of ostensive-inferential communication, but, as I will try to demonstrate, this valuable and important theory faces serious problems when vaguer effects of ostensive behavior need to be described or explained. Since most—and maybe all—Gricean accounts of language are based on the concept of intention, “abnormal” cases such as slips of the tongue, unintended puns, and even misleading transmissions of information are not worthy enough to be studied by pragmatics. The consequence seems to be odd or disappointing: A huge number of verbal manifestations must be paradoxically discarded by theories intending to account for “conversation” in Gricean/cooperative terms. This problem would not affect those trends that deal with non-cooperative forms of language use, for example, some pragmatic analyses of humor (Attardo 1994, 2001, 2006). In addition, there are linguistic theories that intend
AUTHOR’S COPY | AUTORENEXEMPLAR
AUTHOR’S COPY | AUTORENEXEMPLAR
Relevance theory and unintended transmission of information 3 to explain conversation without discarding unintentional verbal manifestations, for example the whole literature of conversation analysis, one of whose cornerstones is the famous work published by Sacks, Schegloff, and Jefferson (1974). Differently from the Theory of Relevance, Neurocognitive Linguistics allows us to assume that the “lapsus linguae” produced by the man who claimed for the necessity of a “prostitute teacher” is especially helpful in giving us clues on how the cognitive-linguistic network operates (Lamb 1999: 181; Lamb 2000: 243). The choice of lexeme can be explained as involving many semantic, lexico-grammatical, and phonological relationships. Cognitive-Philosophical Pragmatics has practically overlapped “intention” and “communication,” thus focusing mainly on the information intentionally transmitted by speakers and writers. Neurocognitive Linguistics helps us to understand that human verbal interaction is much more complex (and interesting) than that. 2. Grice’s influential ideas and the achievements of pragmatics: The case of relevance-theoretic pragmatics Cognitive-Philosophical pragmatics (a.k.a. Anglo-American pragmatics or Gricean pragmatics) aims at accounting for human communication and language understanding. One of its main assumptions is that “pragmatics,” conceived as the study of how contextual factors interact with linguistic meaning in the interpretation of utterances, must be distinguished from “semantics,” the study of core linguistic meaning. In this theoretical framework, there is a semantic and a pragmatic component of language. Many facts can be easily interpreted as instances of human communication and language understanding: A communicator says or writes something, and an interpreter can understand what the communicator said or wrote. Thus, Cognitive-Philosophical pragmatics is (or at least aspires to be) an empirical discipline. However, it has a philosophical origin, and this is what explains part of the name of this linguistic mainstream. The theory of speech acts created by John L. Austin and, particularly, Paul Grice’s theory of meaning led to the real development of the field (Recanati 1987, 1998, 2004a, 2004b). Philosophers interested in ordinary language could not build a bridge between the semantics of formal and natural languages, but Grice suggested that the gap could at least be reduced by distinguishing sentence meaning from speaker’s meaning. Grice also explained that linguistic meanings are composed not only of what is said but also of what is implicated. Schematically, “what is said” can be interpreted as semantic/sentence meaning, while “what is implicated” corresponds to pragmatic/speaker’s meaning. According to
AUTHOR’S COPY | AUTORENEXEMPLAR
AUTHOR’S COPY | AUTORENEXEMPLAR
4 José María Gil Sperber and Wilson (2005: 354) this became the foundation for the majority of modern pragmatics. Different trends within Cognitive-Philosophical pragmatics are grounded on Grice’s theory (1957, 1967, 1982, 1989): (i)
Orthodox Gricean authors conceive that literal meaning must be the point of departure of every analysis. (ii) Neo-Gricean Pragmatics aims at combining an inferential account of communication with a view of language influenced by formal semantics and generative grammar. (iii) The Theory of Relevance decomposes Grice’s intention into informative and communicative intention, and accounts for ostensive-inferential communication. Independently from these particular trends, Cognitive-Philosophical pragmatics has traditionally accepted and encouraged the division of labor between semantics and pragmatics (Stalnaker 1999, Kadmon 2001, Blutner and Zeevat 2003, Asher and Lascarides 2003, S. Davis 1991, Kasher 1998, Horn and Ward 2004). In fact, all the approaches to Cognitive-Philosophical pragmatics are based on these fundamental Gricean ideas: (i)
Speaker’s meaning is a mutually manifest intention. It requires sentence meaning to be conveyed, and it is fulfilled when being recognized by the hearer. (ii) In inferring the speaker’s meaning, the hearer is guided by the expectation that verbal contributions are consistent with rational communicative norms. Grice has suggested that these norms can be included into four classes of “conversational maxims” (Quantity, Quality, Relation, and Manner), which constitute the Cooperative Principle: “. . . in paradigmatic cases, their observance promotes and their violation dispromotes conversational rationality” (Grice 1989: 370). According to Gricean perspective, the speaker makes manifest her intentions and the hearer must recognize those intentions. If the hearer is able to infer the intentions that were made manifest by the speaker, then communication has been successful. Thus, one of the main tasks of pragmatics will be to account for the human capacity of making inferences oriented to identify the speaker’s meaning. For example, Levinson (2006a, 2006b) emphasizes the importance of Gricean intentions and postulates an “interaction engine” that underpins human interaction. Neo-Gricean/formalist scholars root their analysis in the Cooperative Principle and Conversational Maxims (Atlas 2005; Horn 2000, 2004, 2005,
AUTHOR’S COPY | AUTORENEXEMPLAR
AUTHOR’S COPY | AUTORENEXEMPLAR
Relevance theory and unintended transmission of information 5 2007; Levinson 2000) and pay particular attention to generalized implicatures, which are carried by the use of a certain type of words “normally (in the absence of special circumstances)” (Grice 1967: 730). On the other hand, relevance-theoretic authors consider that the formalist approach has little to say about “particularized implicatures” (Sperber and Wilson 2005: 358). In spite of the many theoretical controversies we could find, different trends within Cognitive-Philosophical Pragmatics accept that communication is governed by the expectations about the speaker’s intention. The very idea of intention is also fundamental for the Theory of Relevance. The Communicative Principle of Relevance states that, since the communicator wants her communication to succeed, she indicates by means of her act of communication that she wants the audience to see her utterance as relevant. Relevance-theoretic pragmatics has gone beyond the pioneering work of Grice. For example, it claims that both explicit and implicit communication are inferential, and it maintains that the explicit side of meaning is also worthy of pragmatic attention. Nevertheless, it is assumed that communication is an inferential-ostensive process based on the transmission and recognition of intentions, and it has been repeatedly asserted that it is to Grice that they owe this model of communication. The role of relevance in cognition and communication is accounted for by the two following general principles: Cognitive Principle of Relevance: Human cognition tends to be geared to the maximization of relevance. (ii) Communicative Principle of Relevance: Every act of overt communication conveys a presumption of its own optimal relevance. (i)
The presumption of optimal relevance does not only establish that the utterance is relevant enough to be worth processing, but it also specifies that it is the most relevant utterance which is compatible with the communicator’s intentions. Therefore, relevance-guided comprehension heuristic is determined by communicator’s intention: The addressee will have to follow a path of least effort in constructing an interpretation of the utterance; he will have to resolve ambiguities and referential indeterminacies, to go beyond linguistic meaning, to supply contextual assumptions, and to compute explicatures and implicatures. Finally, the addressee will stop this process when he satisfies his expectations of relevance, which are based on the communicator’s intention. Grice’s distinction between saying and implicating is a natural starting point for examining the semantics-pragmatics distinction (Carston 2002, Wilson and Sperber 2002, Recanati 2004a, b; Szabo 2005; Sperber and Wilson 2005). However, Relevance-theoretic pragmatics assumes that the gap between sentence semantic representation and speaker’s pragmatic meaning is much wider than Grice originally thought, and proposes an inferential system (or
AUTHOR’S COPY | AUTORENEXEMPLAR
AUTHOR’S COPY | AUTORENEXEMPLAR
6 José María Gil uman deductive device) that is rich enough to disambiguate, assign reference, h and compute implicatures. The output of semantics is treated as a highly schematic logical form; pragmatic inferences convert the schematic logical form into a fully propositional form. The semantic component deals with decoded meaning, and the pragmatic component deals with inferred meaning. For example, we may consider (2): (2) John is a computer. The hearer will normally understand that the speaker is not asserting that a particular individual, John, is in fact a particular type of machine. The hearer could normally infer, for instance, that John is an excellent accountant. The relevance-oriented process based on the recognition of the speaker’s intention is summarized in Table 1.
Table 1. A schematic relevance-oriented description of the comprehensive process of an utterance Cognitive task
Specific cognitive task
Linguistic Level of representation
Realizations dealing Example (1)
De-codification (SEMANTICS)
Phonetic decodification
Phonetic Form
´dzon iz ə kəm´pjutər
Semantic decodification
Semantic/Logical Form (in Generative terms)
Disambiguation and reference assignment Identification of propositional attitude Logical form development Logical form enrichment
Propositional form
Inferential comprehension (PRAGMATICS)
Propositional attitude Explicature Implicated premises
Implicated conclusion: Implicature which has been intentionally communicated
CL
[ NP [N John] ] VP [ V being [ FS [ Det a NP [ N computer]]]
Cj C = being a computer j = John The speaker asserted (directly) that John is a computer John is a computer. ‘John is a computer’ is a contradiction. John has been compared to a computer. The comparison is positive. John is an excellent accountant.
AUTHOR’S COPY | AUTORENEXEMPLAR
AUTHOR’S COPY | AUTORENEXEMPLAR
Relevance theory and unintended transmission of information 7 Sperber and Wilson (1995: 9) consider that Generative grammars abstract out the purely linguistic properties of utterances and describe the sentence (the syntactic structure shared by a variety of utterances differing in their non- linguistc properties). Consequently, a good generative grammar accounts satisfactorily for the correspondence between semantic representation and phonetic representation, within the boundaries of sentence. In summary, the study of the semantic representation of sentences belongs to Generative Grammar, while the study of the interpretation of utterances belongs to Pragmatics. In this sense, generative account of grammatical structure (and Universal Grammar) is a theoretical pre-condition for Relevance Theory, which in its turn will characterize communication and cognition (and could also be interpreted as a theory of “Universal Pragmatics”). Table 1 does not include any explicit reference to the syntactic level because syntax is the level accounted for (previously) by Generative Grammar. More specifically, syntax is the nuclear level that produces (i.e., generates) the whole sentence, whose semantic representations and phonetic representations will be treated, respectively, by the “conceptual-intentional system” and the “sensorymotor system.” Those two groups of systems have been considered to be part of “the faculty of language in the broad sense” (Hauser, Fitch, and Chomsky 2002; Fitch, Hauser, and Chomsky 2005). The schematic description of Table 1 should be enriched with some additional commentaries: (a) The implicated premises of the fourth column may have been the result of combining the explicature with some assumptions “stored” in the cognitive environment of the hearer, such as the following: “John is a human being,” “human beings are not computers,” “John is an accountant,” “the speaker answered the question Is John a good accountant?,” “obvious contradictions (such as saying that a person is a machine) may imply a comparison, or some other trope,” “accountants should be efficient accounting,” and “computers can be used to account efficiently.” (b) Relevance-theoretic Pragmatics does not propose a clear borderline between “what is said” and “what is implicated.” However, it does maintain that there is a radical separation between semantics and pragmatics, i.e., between de-codification and inference (Sperber & Wilson 1995, 1998; Carston 1988, 2002; Recanati 1989, 2004a; Wilson & Sperber 2002, Neale 2004). (c) The output of de-codification/semantics is the interpretation of the correspondence between the phonetic representation and the semantic representation. This semantic output is the input for inference/pragmatics. (d) There are “degrees of explicitness” (Sperber & Wilson 1995: 182). The greater the relative contribution of decoding, and the smaller the relative
AUTHOR’S COPY | AUTORENEXEMPLAR
AUTHOR’S COPY | AUTORENEXEMPLAR
8 José María Gil contribution of pragmatic inference, the more explicit an explicature will be (and inversely). Relevance-theoretic pragmatics, which has been interpreted as a prototypical representative of the Cognitive-Philosophical tradition, conceives that comprehension is an online process, that within this process the hearer must find an interpretation of the speaker’s meaning, and that this interpretation must satisfy the presumption of optimal relevance. In order to achieve this, the hearer will have to enrich the decoded sentence meaning at the explicit level, and complement the sentence meaning at the implicit level by supplying contextual assumptions which will combine with the sentence meaning to yield enough conclusions. If the hearer is able to do that, the speaker’s intention will be recognized and the utterance will be relevant. The route followed in dis ambiguating, assigning reference, constructing a context, deriving conclusions, etc., is a path of least effort and its end is the first overall interpretation that satisfies the expectations of relevance. This is the general process that has been exemplified in Table 1, and briefly commented in items (a)–(d). We could consider now some proposals that have been interpreted as achievements of Relevance-theoretic pragmatics. (i)
Table 1 shows some specific cognitive specific tasks such as elaborating hypotheses about explicatures (by developing the linguistically-encoded logical form), or constructing hypotheses about the implicated premises (intended contextual assumptions) and the implicated conclusions (intended contextual implications). Sperber and Wilson claim that these cognitive sub-tasks are not sequentially ordered:
The hearer does not first decode the sentence meaning, then construct an explicature and identify an appropriate context, and then derive a range of implicated conclusions (Sperber and Wilson 2005: 368). Since comprehension is an on-line process, hypotheses about explicatures, implicated premises and implicated conclusions are developed in parallel, against a background of assumptions and expectations that may be checked or elaborated as the utterance unfolds. (ii) The Orthodox Gricean treatment of language use faces serious problems when assuming that language is governed by a norm of literalness. Figurative utterances are analyzed as overt departures from the norm. On the other hand, Table 1 exemplifies the comprehension process of an everyday metaphor. Tropes like that are easily understood without attracting any more attention than purely literal utterances such as “John is an excellent accountant.” This hypothesis has been confirmed by “experimental pragmatics: Reaction-time studies demonstrated that most metaphors take no longer to understand than their literal counterparts (Gibbs 1994;
AUTHOR’S COPY | AUTORENEXEMPLAR
AUTHOR’S COPY | AUTORENEXEMPLAR
Relevance theory and unintended transmission of information 9 Glucksberg 2001; Noveck, Bianco, and Castry 2001; Sperber and Noveck, 2004). Grice’s initial analysis was enlightening, but the idea of a norm of literalness does not seem to be “operationally” plausible (see page 13, after Figure 4). (iii) David Lewis (1979) argued that, in many cases, the standards of precision in force are different from one conversation to another. Austin’s example “France is hexagonal” may be true enough for many contexts, but not true enough for many others. Sperber and Wilson (2005: 376) conclude that Relevance theory is unique in proposing a unified account of literal, vague, loose, and figurative meanings. Since an utterance is a piece of evidence about the speaker’s meaning, the function of the units at the lexical level is not to encode, but to indicate the concepts that are constituents of the speaker’s meaning. Sperber and Wilson consider that words do encode concepts or, at least, semantic features, and they also assume that words are decoded during the comprehension process. Never theless, the output of decoding is simply the point of departure for identifying the explicatures and the implicatures intended by the speaker. There is a mutual adjustment between explicatures and implicatures. The decoded content helps to identify the inferences that make the utterance relevant as expected, and it is readjusted in order to warrant just those inferences that contribute to the relevance of the utterance as a whole (Carston 2002; Wilson and Sperber 2002). Traditional taxonomies do not seem to be “operationally” plausible, because the comprehension process does not involve interpretations such as literal, vague, loose, metaphorical, etc. Relevance-theoretic pragmatics suggests that these concepts make reference to fuzzy zones on a continuum; and the most important fact is that they can be useful in metalinguistic analyses, but play no role in a realistic theory of language use. (In section 2, which introduces Neurocognitive Linguistics and its relational network theory, it will be made more explicit how the “realistic” is interpreted). (iv) Relevance-theoretic pragmatics can account for linguistic constructions that contribute to aspects of a speaker’s meaning different from explicit truth-conditional content, or constructions that encode aspects of meaning that are not plausibly analyzed in conceptual terms. For example, this may include illocutionary force indicators, presupposition triggers, indexicals and demonstratives, focusing devices, parentheticals, discourse connectives, argumentative operators, prosody, interjections, etc. (Sperber and Wilson 2005: 378–379). These non-truth-conditional aspects of speaker’s meaning are interpreted as higher-level explicatures elaborated by development of encoded schematic sentence meanings. For instance, the semantic function of deictic expressions serves to indicate that a referent is required and to restrict the search to a certain class of
AUTHOR’S COPY | AUTORENEXEMPLAR
AUTHOR’S COPY | AUTORENEXEMPLAR
10 José María Gil candidates. The encoded meaning of a deictic expression such as “here” is only a clue for the speaker’s meaning, which is recovered, as always, by mutual adjustment of explicatures and implicatures in the search for optimal relevance. It is not the case that all linguistic meaning must be either conceptual or truth-conditional: The only requirement on linguistic meaning is to guide the hearer toward the speaker’s meaning by indicating the direction in which relevance is to be sought. (v) Truth-value is also a pragmatic issue, because it depends on referential conditions. Within this framework, reference assignment belongs to the set of inferential sub-tasks, and the cognitive/linguistic units by means of which we make reference to “the real world” are explicatures and implicatures (and they are also result of the pragmatic process of inference). (vi) Relevance Theory is able to treat meaning problems that had been discarded by traditional formal approaches. Austin (1962) claimed that, in order to understand the meaning of a sentence, it was not enough to understand its truth condition. Grice postulated that the intentional speaker’s meaning was amenable to philosophical and scientific treatment. After many years of research, Relevance Theory states that it is apparent that the gap between sentence meaning and speaker’s meaning is wide. Currently, it is argued that linguistic/semantic meanings are mental representations playing a role at an intermediate stage in the comprehension process: They are not consciously entertained. On the other hand, speakers’ meanings are salient objects in the psychological process of verbal comprehension. Although it has a philosophical origin, Relevance Theory intends to be an empirical science and faces the challenge of explaining “how the closed formal system of language provides effective pieces of evidence, which, combined with further evidence from an open and dynamic context, enable hearers to infer speakers’ meanings” (Sperber and Wilson 2005: 383; italics are mine). Cognitive psychology will help by providing methods such as modeling of mental processes, experimental tests, evolutionary insights, studies of language pathologies like aphasias, and analyses of communication pathologies such as dysphonia (Paradis 1998) or autism. Although it stems from the Gricean tradition, Relevance-theoretic pragmatics has developed a more comprehensive theory of language use. It maintains that both explicit and implicit communication are inferential, and verbal comprehension is conceived as a psychological process governed by the search or relevance, a rational and anthropological principle that seems to underlie human biology and human culture. Within this theoretical framework, the concept of intention is highly important: The hearer’s tasks are aimed at discovering the speaker’s intended meaning in order to satisfy the presumption of
AUTHOR’S COPY | AUTORENEXEMPLAR
AUTHOR’S COPY | AUTORENEXEMPLAR
Relevance theory and unintended transmission of information 11 relevance. However, Sperber and Wilson (1995: 201) have admitted that pragmatics should not be concerned purely with the recovery of an enumerable set of assumptions individually intended by the speaker. For example, poetic effects in verbal comprehension can be defined as the very weak implicatures that are mainly responsible of hearer’s (or reader’s) interpretation. But this hypothesis is odd. If we accept it, Relevance-theoretic pragmatics postulates a contradiction: There are interpretations that are not consistent with the speaker’s intention (and the Communicative Principle of Relevance). In the third section, after introducing relational network theory, we will consider some examples of verbal productions and comprehension processes which cannot be properly analyzed in terms of the Theory of Relevance. 3. Neurocognitive linguistics and relational network theory According to Neurocognitive Linguistics, the linguistic system is a particular brain system richly and strongly connected with other brain systems. In fact, the term “language” is conceived as a simple label that we have put into use in order to talk about a particular configuration of interconnected subsystems (phonological recognition, phonological production, lexical systems, morphology, semantic of concrete nouns, etc.). We like to think about those different systems as if they were unitary. Within this context, the linguistic system of the individual is conceived as a real biological system that has the form of highly complex set of networks (Lamb 2000). As we will see later, neurological plausibility is explicitly stated as one of the requisites of the theory. Since the linguistic system is a network, there must be some type of notation accounting for this fact. Consequently, relational network theory will provide concrete diagrams that depict the actual linguistic system of an individual. For example, Figure 1 can help to understand one of the main hypotheses of Neurocognitive Linguistics: The linguistic system is a network of relationships. Those networks are integrated by nections and connections. Nections are the fundamental modules of neurocognitive relational networks: They have a central line connecting two nodes, one with upward branching nodes, and the other one with downward branching nodes. Connections are the lines linking nections. Dealing with neurological plausibility, nections and connections will be implemented, respectively, as cortical columns and neural fibers (Lamb 2005, Lamb 2006). Figure 1, which is based on a figure provided by Lamb (1974: 147), represents a neurocognitive interpretation of the English Subject and other syntactic functions. Like every relational network, it serves to illustrate that the linguistic system can be conceived as a real network, that is to say, a type of structure that branches in both directions, so it is possible to have paths that first branch
AUTHOR’S COPY | AUTORENEXEMPLAR
AUTHOR’S COPY | AUTORENEXEMPLAR
12 José María Gil
Figure 1. Some nections and connections related to Subject, Object, and Noun Phrase
and then come together. The network system consists of lines and nodes, and each node is a point at which lines intersect. There is a basic distinction between OR and AND nodes, which specify, respectively, paradigmatic and syntagmatic relationships (see Figure 4). In the case of Figure 1, the upward direction within the syntax leads to different syntactic functions: The English Noun Phrase has an upward OR node that relates noun phrases to the different functions they have, like Subject and Object. There is a downward OR node that connects to alternative linguistic units that can have the same function: For example, pronouns or Noun Phrases can play the role of Subject or Object. Of course, there are also AND (syntagmatic) relationships within syntactic structures. The downward AND node is used, in Figure 1, for a combination of syntactic units performing a specific function. Thus, a Noun Phrase connects downward in an ordered AND relationship to a structure integrated by a noun that is preceded by a determiner and an adjective, and followed by a prepositional phrase, like in the little house on the prairie. The ordered downward ORs corresponding to Determiner, Adjective, and Prepositional Phrase indicate that the default choice is simply nothing (represented by a small circle), an example being the noun phrase girls in the clause girls just want to have fun. The upward unordered AND can be used, for example, to indicate the connection between a lexicogrammatical nection, like
AUTHOR’S COPY | AUTORENEXEMPLAR
AUTHOR’S COPY | AUTORENEXEMPLAR
Relevance theory and unintended transmission of information 13
Figure 2. Example of upward unordered AND
Figure 3. The lexeme hard as represented in the system Note: (1) Nections for distinctive features are connected to nections for phonemes. (2) Nections for phonemes are connected to nections for morphemes and lexemes. (3) Nections for morphemes and lexemes are connected to meanings: In this particular sense, it is said that “words” do not encode meanings, but are connected to them.
the one for is, and the nections that specify its meaning: 3° person singular, BE, and PRESENT (Figure 2). Relational networks, like Figure 3, enable us to depict clearly the distinction between phoneme features, phonemes, lexemes and meanings. In relational network theory, lexicogrammatical nections are connected to semantic and phonological nections; there are no such things as semantic and phonological entries for lexical units.
AUTHOR’S COPY | AUTORENEXEMPLAR
AUTHOR’S COPY | AUTORENEXEMPLAR
14 José María Gil Downward Unordered “AND” Downward activation from a goes to b and c. Upward activation from b and c goes to a.
Exemplified and indicated in Figure 3.
Upward activation from a goes to b and c. Downward activation from b and c goes to a.
Exemplified and indicated in Figure 2.
Downward activation from a goes to b and later to c. Upward activation from b and later from c goes to a.
Exemplified in Figures 1 and 3. Indicated in Figure 1.
Downward activation from a goes to b and [sic] c. Upward activation from b or c goes to a.
Exemplified in Figures 1 and 3. Indicated in Figure 1.
Upward activation from a goes to b and c. Downward activation from b or c goes to a.
Exemplified in Figures 1 and 3. Indicated in Figure 1.
Downward activation from a goes to b if possible, otherwise to c. Upward activation from b or c goes to a.
Exemplified and indicated in Figure 1.
Upward Unordered “AND”
Downward Unordered “AND”
Downward Unordered “OR”
Upward Unordered “OR”
Downward Ordered “OR”
Figure 4. Nodes of relational networks in abstract notation (adapted from Lamb 1999: 67)
Figures 1, 2, and 3 provide examples of all the types of nodes used in the abstract notation of relational networks, also known as “compact notation” (Lamb 1999: 67). The nodes differ from one another according to three parameters of comparison: (i) upward vs. downward orientation, (ii) and vs. or, (iii) ordered vs. unordered. Figure 4 shows that we have six types of nodes in ab-
AUTHOR’S COPY | AUTORENEXEMPLAR
AUTHOR’S COPY | AUTORENEXEMPLAR
Relevance theory and unintended transmission of information 15 stract notation. It must be emphasized that activations are bidirectional here. (There is another type of notation where activating lines are unidirectional, but there is not enough space to talk about this “narrow notation” here.) Lamb proposes a definite meaning to the term “realistic” in the context of his neurocognitive research (Lamb 1999, 2000, 2005, 2006). A “realistic” theory of language should go beyond the analysis of the products of verbal behavior (i.e., texts), and should account for the linguistic system in relation to actual human beings. With a view to doing that, a realistic linguistic theory will have to satisfy the following three requirements (Lamb 1999: 293–294): Operational plausibility: A realistic linguistic theory has to provide a plausible account of how the linguistic system can be put into operation in real time to produce and understand speech. (ii) Developmental plausibility: A realistic linguistic theory needs to be amenable to a plausible account of how the linguistic system can be learned by children. (iii) Neurological plausibility: A realistic linguistic theory has to be compatible with what is known about the brain from neurosciences. (i)
We could suggest that a linguistic theory is unrealistic if it is disconfirmed by some kind of evidence. For example, if a particular linguistic theory presents hypotheses about language learning and language structure that are disconfirmed by neurological evidence, this linguistic theory will be reasonably considered to be unrealistic. This would be the case for a theory proposing that an individual’s linguistic and semological systems consist of arrangements of symbols, being “semology” the whole system of meanings within the individual’s cognitive system. This hypothesis is disconfirmed by the following facts (Lamb 2005): (i) it requires a device in the brain that can read information in symbolic form (and our brains do not have such a device); (ii) the brain does not store symbols, like a computer does; (iii) the process of interpreting symbols requires additional devices: storage for the symbols, a buffer in which to store the input item while the process of recognition is going on, a device to perform comparisons (and, again, our brains do not have such additional devices). Against the hypothesis that our brain stores symbols, a more realistic alternative is to assume that the internal structure does not have symbolic representations of phonemes, morphemes, or lexemes, but the means for producing such forms in oral or written texts. Relational network theory is attractive from a neurological point of view because it seems to be compatible with neurological evidence. Neuroscience research has shown that the cerebral cortex is a network, and that learning develops as strengthening of connections. Basic processes involved in text comprehension operate directly in the network “as patterns of activation traveling
AUTHOR’S COPY | AUTORENEXEMPLAR
AUTHOR’S COPY | AUTORENEXEMPLAR
16 José María Gil the pathways formed by its lines and nodes” (Lamb 2005: 157). Linguistic and semological information is not stored as symbolic representations, but it is in the connections. There is a good amount of neurological evidence for relational networks. However, there is no direct experimental evidence for some of their features because of the following reasons: (i)
Brain images are too rough for the study of microscopic levels (Cherchi 2000, Lamb 2004b). (ii) The experiments with living brain tissue of animals are not done with humans for obvious ethical reasons. (iii) The experiments with living brain tissue of animals deal with visual, auditory, and somatosensory perception of cats and monkeys (Hubel and Wiesel 1962, 1968, 1977; Mountcastle 1997, 1998), and these animals do not perform linguistic processing. On the other hand, there is a good amount of relevant indirect evidence for the neurological plausibility of relational network theory. For example, Hubel and Wiesel (1962, 1968, 1977) discovered that visual perception in cats and monkeys works in the ways that would be predicted by the relational network model, and the nections of visual network are implemented as cortical columns: “The nodes are organized in a hierarchical network in which each successive layer integrates features from the next lower layer and sends activation to higher layers” (Lamb 2005: 168). Mountcastle (1998: 165) suggests the following hypothesis: “the mini column is the smallest processing unit of the neocortex,” and he also claims that, “every cellular study of the auditory cortex in cat and monkey has provided direct evidence for its columnar organization” (1998: 181). For example, a nerve-regeneration experiment in the monkey provides evidence for columnar organization of the somatic sensory cortex. A recording microelectrode was passed nearly parallel to the pial surface of the cortex of the postcentral somatic sensory cortex, through a region of neurons with the same modality properties. Neurons in adjacent minicolumns are related to adjoining and overlapping peripheral receptive fields, and the transitions between minicolumns pass unnoticed. Results obtained in the same animal in a similar experiment after section and resuture of the contralateral medial nerve showed a misdirection of the regenerating bundle of nerve fibers, innervating then the glabrous skin of the hand. Sudden displacements of receptive fields, which occur at intervals of 50–60 microns, reveal the minicolumns and their transverse size (Mountcastle 1997: 708, 1998: 173). Since speech perception is a higher-level perception process, it is permissible to suggest the following extrapolation: Each node (or nection) in the neurocognitive system of an individual can be implemented as a cortical column.
AUTHOR’S COPY | AUTORENEXEMPLAR
AUTHOR’S COPY | AUTORENEXEMPLAR
Relevance theory and unintended transmission of information 17 Within the linguistic system, every nection/cortical column has a highly specific function. For example, there may be a nection/cortical column corresponding to the lexeme hard and to the sememe HARD as they are represented in Figure 3. Now, we have that the relational network model requires (before considering its neurological plausibility) the following types of connectivity among its nections, and the following types of properties for its connections (Lamb 2005: 170): 1. 2.
Connections can have varying strengths. Connections are strengthened through successful use (the learning process). 3. Connections of given strength carry varying degrees of activation. 4. Nections have varying thresholds of activation. 5. The threshold of a nection can vary over time (part of the learning process). 6. Connections are of two types: excitatory and inhibitory. 7. Excitatory connections are bidirectional, feed-forward, and feed- backward. 8. Excitatory connections can be either local or distant. 9. Inhibitory connections are local only. 10. Inhibitory connections can connect either to a nection or a line; the blocking element attaches to a line. 11. In early stages (pre-learning) most connections are very weak (latent). 12. A nection (at least some nections) must contain an internal wait (delay) element, needed for sequencing, for example, of the part of a syllable or of the constituents of a construction. The examination of evidence shows that minicolumns and their interconnections have every one of these properties. For example, the internal delay element (Number 12 in the above list) is implemented by means of axon fibers that branch off from the axons of pyramidal cells within a column and connect vertically to other cells in the same column. “[F]rom layer VI they project upwards and from upper layers downward. This circulating activation among the pyramidal cells of a column keeps activation alive until it is turned off by inhibitory neurons with axons extending vertically within the same column. Such inhibitory cells are called double basket cells” (Lamb 2005: 170). There are also relevant considerations about the number of minicolumns that an individual would need in order to represent linguistic information. For example, when estimating the huge number of minicolumns in Wernicke’s area, Lamb (2005: 172) suggests that there could be approximately 2,800,000 minicolumns in that area. This number could allow an individual to represent all the information needed for phonological perception.
AUTHOR’S COPY | AUTORENEXEMPLAR
AUTHOR’S COPY | AUTORENEXEMPLAR
18 José María Gil On the basis of previous remarks, we could provide an argument for the neurological plausibility of relational networks: Argument for the neurological plausibility of relational networks i. Nections represented in relational networks are implemented as cortical columns. ii. Connections represented in relational networks are implemented as neural fibers. iii. Cortical columns and neural fibers integrate real cortical connections. iv. Therefore, relational networks represent real cortical connections. The following are, finally, the main characteristics and functions of neurocognitive relational networks (NRN): (i)
NRN account for the brain systems, which allow an individual to produce the utterances. (ii) NRN can be used to represent any relevant part of the actual linguistic and semological systems of a real speaker or hearer. (iii) NRN demonstrate a clear distinction between phonological, lexicogrammatical and semantic nodes, and exhibit the connections among them. (iv) NRN account for automatic and parallel processing in linguistic production and comprehension. (v) NRN make explicit the semantic information that governs lexicogrammatical choices: Semantic nections lead downward to lexicogrammatical ones. (vi) NRN are purely relational. In Hjelmslev’s terms, NRN free linguistic science from the metaphysical hypothesis that objects and symbols are something different from relationships (Hjelmslev 1943). Since symbols are not part of linguistic and semological systems, the inscriptions beside relational networks are just labels.
4. The limitations imposed by the very concept of intention: Some examples that cannot be accounted for in terms of relevance theory The very concept of intention has been repeatedly criticized. For example, Jeff Verschueren called for “a pragmatic return to meaning in its full complexity, allowing for interacting forces of language production and interpretation” (1999: 48). In the case of Sociocultural-Interactional Pragmatics (another important approach that cannot be analyzed here) the concept of “intention” does not only receive little attention, but it has also been regarded as a “culture-specific” no-
AUTHOR’S COPY | AUTORENEXEMPLAR
AUTHOR’S COPY | AUTORENEXEMPLAR
Relevance theory and unintended transmission of information 19 tion (Danziger 2006; Duranti, 2006; Richland 2006). Recently, there have been contributions that posit that Gricean intentions are not constitutive to verbal interaction (Arundale 2008) and that verbal information can be transmitted without communicative intention (Németh T. 2008). Within the Cognitive-Philosophical perspective on pragmatics, the role of the speaker’s intention in the communicative process has been ardently debated (W. Davis 1998, 2007, 2008; Gibbs 1999, 2001; Green 2007, 2008; Jaszczolt 2005, 2006; Keysar 2007; Saul 2001; Thompson 2008). In his article introducing a debate on intention in Intercultural Pragmatics, Michael Haugh (2008: 102) says that, while there is substantial (if not overwhelming) evidence against the placement of Gricean intentions at the center of theorizing in pragmatics, “there remains a need to account for the cognition that underlies interaction.” Sperber and Wilson (2005: 358) declare that that formal/Neo-Gricean pragmatics narrows the domain of pragmatics research because it has little to say about particularized implicatures. It is possible to suggest the following hypothesis: Most trends of Philosophical-Cognitive Pragmatics have narrowed the domain of research because they have overestimated the importance of “intention.” In fact, most of current pragmatics has little to say about many valuable cases that reveal very important data about cognitive processes underlying interaction. Thus, unintended puns and slips of the tongue are not interpreted as cases of verbal communication. For example, after hearing that the sixth-grade teacher was going to be absent for a three-month period, one of the fathers attending a meeting said: (1) We need a prostitute teacher. I will consider that (1) can be interpreted (at least in one of its many aspects) as a typical example of “unintended transmission of information.” This phenomenon could be also called “unintended communication” or even “unintentional meaning,” but these particular labels will not be used here in order to avoid futile terminological problems; within Relevance Theory, communication and meaning are intentional. In fact, communication has to be defined on the basis of communicative intention and relevance. The idea of “unintended transmission of information” presented here requires a definition of what information unintentionally transmitted is: (i) Information unintentionally transmitted: Meaning that has been interpreted by the audience on the basis of the communicator’s utterance/s, but considered by the audience as not predictable or not controllable for the communicator. Making use now of Sperber and Wilson’s words: Information unintentionally transmitted, which actually modifies the cognitive environment of the
AUTHOR’S COPY | AUTORENEXEMPLAR
AUTHOR’S COPY | AUTORENEXEMPLAR
20 José María Gil a udience, is some meaning that has been transmitted by the communicator’s utterance, but not having been made manifest or more manifest to the audience by the communicator herself. Sperber and Wilson (1995: 57) recognize that “something can be communicated without being strictly speaking meant by the communicator or the communicator’s behaviour,” i.e., some information or some meaning can be transmitted and interpreted independently from the communicative intention. I do think that this hypothesis is true, but I also think that Relevance Theory cannot justify it. As an alternative to Relevance Theory, I shall argue that this “information unintentionally transmitted” (which of course could be expressed in terms of a set of assumptions) is represented in the neurocognitive system of an individual, and that relational networks can account for it. On the other hand, the set of assumptions describing such information is just one of the products of the neurocognitive system of the individual (where there are no things such as symbols, arrangements of symbols, assumptions, or sets of assumptions). The father in example (1) said “prostitute” instead of “substitute,” probably conditioned by the rhyme and other associations. This “lapsus linguae” was produced spontaneously; the evidence for this is that the producer of (1) was surprised both when his hearers laughed and when the slip of the tongue was pointed out. According to canonical Cognitive-Philosophical pragmatics, such an example is not an instance of verbal communication at all, simply because the speaker did not have the intention to mean anything about any prostitute. The producer of (1) transmitted some information about his feelings or thoughts, but this transmission is linguistically irrelevant for pragmatics: It could be compared with the information that is unintentionally transmitted by the trembling of the voice. An instance of “lapsus linguae” could not be explained in terms of the Cooperative Principle, simply because the meaning evoked by the utterance is very different from the speaker’s meaning. Regarding the Theory of Relevance, and although it occurs during ostensive behavior, the Freudian slip produced in (1) is not ostensive; it is not the result of communicative intention. In Sperber and Wilson’s words, the producer of (1) did not want to make mutually manifest to audience and himself that he had the intention to inform the necessity of a prostitute. Sperber and Wilson (1995: 201) admit that pragmatics should not be concerned exclusively with the recovery of an enumerable set of assumptions individually intended by the speaker. However, the comprehension of “unintended transmission of information” is not guided by the Communicative Principle of Relevance. This principle establishes that every act of overt communication conveys a presumption of its own optimal relevance, i.e., the utterance is not only relevant enough to be worth processing, but it is also the most relevant one that is compatible with the communicator’s intentions. It is apparent
AUTHOR’S COPY | AUTORENEXEMPLAR
AUTHOR’S COPY | AUTORENEXEMPLAR
Relevance theory and unintended transmission of information 21 that the hearers of (1) will be able to draw some inferences and to elaborate some hypotheses about the speaker’s mental representations. For example, some of them could infer that the speaker has unconscious but strong representations about sexual desire. This interpretation seems to reveal the organization of some part of the cognitive system of a particular individual. Phonological associations favored semantic associations and, unintentionally, the speaker transmitted some of his semantic representations. It is also relevant here that “hire a prostitute” is a more frequently occurring sequence than “hire a sub stitute”; for example, there are 105,000 hits for the first on a Google search, and only 46,000 for the latter. Thus, we can suppose that “hire a prostitute” is on its way to becoming lexicalized, which also influences the situation. It is not unreasonable to infer that the speaker had some (at least vague and probably “unconscious”) representation/s about sex when he said, “We need a prostitute teacher” instead of saying, “We need a substitute teacher.” However, the speaker’s meaning or the speaker’s communicative intention do not indicate where relevance is to be sought. The hearers found their own way, independently of this father’s intentions. There are no explicatures and no implicatures when we infer, on the basis of the speaker’s utterance, that some mental representations (which have not been ostensively and intentionally communicated) are present in the speaker’s cognitive system. Some of these mental representations could be represented in the relational network of Figure 5, which helps us understand that the situation can be described by recognizing that the individual’s cognitive and linguistic systems have the form of a network. The notation used in the upper part of this figure is a simplified version of the abstract notation (the lines are still bidirectional). Here, each nection is represented by one circle and connections are shown as single lines. By means of this simplified notation, the diagram shows that semantic nections are connected, but it makes no attempt to show specific details of the semantic level. This relational network also represents roughly both (i) what could be interpreted as intentional meaning (with a gray thick line), and (ii) what is interpreted as “unintended transmission of information” (the black thick line). Since accounts of language in Gricean pragmatics are based on the concept of intention, supposedly abnormal cases such as slips of the tongue, unintended puns, and even misleading transmissions of information have not been considered worthy of study. (Below, Figure 8 aims at representing neurocognitive structures upon the bases of which we could talk about “misleading transmission of information.”) The consequence of this seems odd and disappointing— a huge number of verbal manifestations are paradoxically discarded by some linguistic theories that claim to account for conversation or communication. On the other hand, Neurocognitive Linguistics allows us to assume that the lapsus linguae produced by the man who called for a prostitute teacher is especially helpful in giving us clues to how the linguistic network operates (Lamb
AUTHOR’S COPY | AUTORENEXEMPLAR
AUTHOR’S COPY | AUTORENEXEMPLAR
22 José María Gil
Figure 5. Slip of the tongue involving prostitute and substitute
1999: 181, 2000: 243). The choice of lexeme can be explained as involving many semological, lexical, and phonological relationships. In summary, at least part of the meaning transmitted (or evoked) by example (1) is not intentional; namely, the father in that meeting did not want to make manifest for himself and his audience that the school needed to hire a prostitute (instead of a substitute). This “information unintentionally transmitted” deserves to be studied simply because it is one of the many phenomena that appear in conversation, i.e., in verbal interaction. Within this context, it must be emphasized that the study of “information unintentionally transmitted” seems to provide support for the network model. We can thus suggest that the nection corresponding to prostitute was more strongly activated that the one corresponding to substitute because the former received more activation from the semantic level. The case is very interesting, because the nections prostitute and substitute are both connected to the pho nological nections /sti/ and /tut/ (which come after phonological constituents /pro/ in prostitute and /sub/ in substitute). The example can be clearly explained in terms of neurocognitive relational network theory, since excitatory connections not only can have varying strengths but are also bidirectional (feed-forward and feed-backward) (Lamb 2005: 170). The excitation coming from semantic nections “sex” and “prostitute” was stronger than the one coming from “school,” “substitute,” and “school,” just because (at that moment) these representations were being more strongly activated in the semological system of the speaker. We could say that certain meanings in the neurocognitive system of the speaker were stronger than the meanings evoked by conversation, i.e., the activation of “prostitute” was (at that moment) stronger than the
AUTHOR’S COPY | AUTORENEXEMPLAR
AUTHOR’S COPY | AUTORENEXEMPLAR
Relevance theory and unintended transmission of information 23 activation of “substitute,” which contributes to explain the selection of lexeme prostitute instead of substitute. It must be admitted that relational network theory cannot explain why the nections corresponding to “sex” and “prostitute” were more strongly activated. As a scientific theory, relational network theory does not extend quite far enough. Similarly, philosophical and linguistic theories that account for intentional meaning cannot explain, for instance, why the father who promises sincerely to buy a chocolate does have the firm intention to accomplish his promise (buying a chocolate). In addition, within the neurocognitive approach it is admitted that the linguistic system can change at any time, and it can also be expected to change every time it is used. It may be also signaled that cognitive and neurological existence does not necessarily entail placement in specific loci in the brain. For example, intended and unintended meanings, or unconscious and conscious connections, have neurological existence as neurocognitive functions. In fact, the unconscious and the unintended meanings can be interpreted as neurocognitive functions that are realized by means of actual neurocognitive structures, namely the connections between lexemes and meanings as well as between meanings. Much of this has been suggested in the relational analysis of slips of the tongue and unintended puns (Dell 1979; Dell and Reich 1977, 1980a, 1980b; Reich 1985; Lamb 1999, 2000, 2005). This line of analysis could provide linguistic support to empirical psychoanalytical research. There are many other relevant examples that can be provided and consistently analyzed in terms of relational network theory. This framework accounts for linguistic information in terms of connectivity and relationships, and it also enables us to explain some problematic cases that Gricean Pragmatics has traditionally overlooked. Consider another example. In most Hispanic countries, soccer is an effervescent popular passion. Some time ago, the famous former soccer player Diego Maradona was chosen as the coach of the national team of Argentina. Many fans and journalists are Maradona worshippers, but many others are not. Consequently, there were ardent controversies in the media. During one of them, a young journalist (who had insistently made manifest his admiration for Maradona) exclaimed the following: (3) Maradona aspira a todo. [Maradona has great aspirations]. The original example constitutes an unintended pun. Depending on the context, the Spanish verb aspira means ‘aspire,’ ‘vacuum,’ or ‘inhale.’ This third option is critical to the pun because it is well known that Maradona used to have serious problems with his addiction to cocaine, a drug that is aspirada ‘inhaled.’
AUTHOR’S COPY | AUTORENEXEMPLAR
AUTHOR’S COPY | AUTORENEXEMPLAR
24 José María Gil
Figure 6. Unintended pun involving the utterance “Maradona aspira a todo”
There are some clues that help us to accept that this example is an unintended pun (or another Freudian slip). The young journalist who said (3) was praising Maradona’s work. Two members of his audience laughed immediately after the utterance and made him notice that the word aspira could sound indeseablemente irónica ‘inconveniently ironic.’ There is another important fact that allows us to interpret the pun as unintended: The journalist was not only visibly embarrassed and annoyed after being made aware of the pun, but also said that that interpretation was una inaceptable falta de respeto para Maradona (‘an unacceptable lack of respect for Maradona’). Relational network theory can account for the information in the lexico-grammatical and semantic systems of the young man who uttered (3). Figure 6 provides a possible organization, which depicts both the basis of what will be interpreted as “unintended transmission of information” (black thick line) and what could be interpreted as “intentional meaning” (gray thick line). Lamb suggests that the object of neurocognitive linguistics is the individual’s linguistic system. Of course, this proposition does not exclude the possibility of studying verbal interaction. In this case, the hearers draw inferences about the organization of the speaker’s cognitive and linguistic systems. Thus, the relations characterized in Figure 6 account for the speaker’s systems, but they also make reference to the interpretation of the hearers. In this sense, neurocognitive linguistics is also “interactive, . . . [and] you need a cognitive approach to fully understand human interaction” (Lamb 2004c: 437). It must be emphasized here that the study of verbal interactions is perfectly possible (and even necessary) for the neurocognitive perspective. The linguistic system of every person is different from those of every other person. Thus, there is no possibility of perfect communication through language. However, there are remarkable similarities between the systems of any two persons of different parts of the world, allowing at least some degree of interaction even among speakers of different languages. The linguistic subsystems (like phonology,
AUTHOR’S COPY | AUTORENEXEMPLAR
AUTHOR’S COPY | AUTORENEXEMPLAR
Relevance theory and unintended transmission of information 25 lexico-grammar, and semantics) have their basic structural properties in common with other cognitive subsystems (like somato-sensory perception, audition, and vision), including hierarchical organization and similar nection structures and network operations (Lamb 1999: 376). Unintended pun comprehension can be accounted for in relational network terms. The hearers interpret that the cognitive system of the young journalist includes representations that are not consistent with those he wanted to transmit intentionally; namely, there are representations about the link between Maradona and drug inhalation. The type of global comprehension illustrated in Figures 5 and 6 is actually performed by the interpreters on the basis of utterances such as (1) and (3). However, those inferential processes could never be characterized if the hearers (only) wanted to find an interpretation of the speaker’s meaning satisfying the presumption of optimal relevance. It has to be mentioned that relational network theory is also useful to account for unintentional ambiguities such as (4), which could hardly be interpreted as a Freudian slip. (4) They are flying objects. Relational network theory also allows us to describe intentional ambiguities, where the speaker “flouts” and “exploits” one of the sub-maxims of Manner (Grice 1967: 725). Let us consider the example of the British General who captured the province of Sind, at that moment in British India, and sent back the message (5) (Grice 1967: 729). (5) I have Sind. It must be acknowledged again that intentional meaning and information that is unintentionally transmitted generally (maybe always) exist together. Figure 5 and Figure 6 account for this fact. In other words, many times (maybe always) the audience’s “cognitive environment” can be modified or enriched both by the information that has been made intentionally manifest by the communicator, and by the information that was recognized by the audience independently (and sometimes against) such intentional meaning. In this sense, what the communicator says can include both what s/he means, and what s/he does not mean. Complementarily, what the audience interprets has its roots both in intentional meaning and in information unintentionally transmitted. Figure 7 accounts for the representation of intentional meaning in relational network terms. The key factor here is that relational network theory allows us to account for both intended and unintended forms of verbal interaction. On the contrary, Gricean Pragmatics does not seem to account for utterance comprehension independently from the speaker’s intention.
AUTHOR’S COPY | AUTORENEXEMPLAR
AUTHOR’S COPY | AUTORENEXEMPLAR
26 José María Gil
Figure 7. Intentional ambiguity involving sinned and Sind
Relational network theory could be also useful to explain other complex cases that do not fit the very concept of intention. An instance of an unintentional contradiction could be interpreted as a signal of the speaker’s cognitive system. Those kinds of contradictions are very easy to find in declarations or speeches delivered by show-business stars and politicians. For example, a well-known actress said, in different scenes of the same TV show, that a certain colleague of hers was “despicable” and “an honorable man.” Previously, we said that relational network theory could account for both intentional and unintentional samples of verbal interaction. We can go further now: This theory describes satisfactorily not only the individual’s cognitive structure that is able to evoke (intentionally or unintentionally) some meanings, but also the individual’s cognitive structure that is capable of interpreting the meanings that have been evoked by the speaker. There are many other cases of “unintentional communication”: accent (which reveals where a speaker is from), voice quality (which reveals whether male or female; adult or child; whether agitated, anxious, fearful, jovial, etc.); choice of words may indicate the level of education. Many of these aspects have been considered by sociolinguistics. Misunderstandings have also been extensively discussed within relevancetheoretic pragmatics (Bou-Franch 2002; Ivaskó and Németh T. 2002; Mirecki 2004, 2005; Yus 1998, 1999a, 1999b). Patricia Bou-Franch analyzes an interesting example within the context of a longer interaction. A former secretary
AUTHOR’S COPY | AUTORENEXEMPLAR
AUTHOR’S COPY | AUTORENEXEMPLAR
Relevance theory and unintended transmission of information 27 identified as TM is now a top executive who has her own personal assistant, but she is not aware of her new job. After having arrived in the office, TM interpreted that another lady (identified as AB) is her new boss, but in fact AB is TM’s personal assistant. (6) AB: How about some coffee? TM: Sure just tell me where anMisunderstandings can also occur when hearers do not interpret the content of the speakers’ informative intention. In this case, the communicators’ communicative intention is fulfilled, but their informative intention is not. In example (6), since TM did not activate the crucial contextual assumptions that are consistent with the communicative principle of relevance, the response uttered by her is an example of misunderstanding. (TM makes manifest that she will be the one in charge to serve the coffee, but she has been offered coffee by AB). Due to that lack of activation, there are other misunderstandings that have consequences throughout the whole interaction. The examples analyzed within the relevance-oriented approach are conceived as misunderstandings precisely because the speaker’s communicative intention has not been identified by the addressee. But this analysis does not mention two facts that should make misunderstandings particularly interesting for language use study: (i) they are samples of the cognitive system of a particular individual at the moment of the interaction; (ii) they are cases of unintentional transmission of information (and consequently, incompatible with the speaker’s intention). Our linguistic systems are able to produce (and understand) utterances that evoke many types of meanings. Some of these meanings are intentionally evoked, with mutual manifestness. But other meanings are not intentionally evoked, like those of unintended puns, slips of the tongue, unintended ambiguities and contradictions, etc. Within this perspective, we would like to suggest the possibility of studying a particular type of verbal products that could also be relevant for language use theory—misleading utterances. For example, when hearing a politician, an interpreter can believe that the politician is lying. Therefore, this interpreter draws many inferences about the speaker’s cognitive system. This seems to be a very common case in everyday life. However, Cognitive-Philosophical pragmatics has little or nothing to say about many ordinary interactions in which the hearer believes that the speaker is lying. According to the canonical description given by Grice, the speaker is “misleading” when she “quietly and unostentatiously” violates a maxim and thus “dispromotes” communication or conversational rationality. According to Sperber and Wilson, the speaker is misleading when she knows that her utterance is not the most relevant she could have uttered; in this case, the speaker unostentatiously contradicts the principle of optimal relevance, which will confuse the hearer’s interpretation. As a result, misleading
AUTHOR’S COPY | AUTORENEXEMPLAR
AUTHOR’S COPY | AUTORENEXEMPLAR
28 José María Gil utterances are not genuine cases of (intentional) communication, and therefore have been practically excluded from the scope of pragmatics. In fact, the theoretical framework developed by Relevance Theory cannot account for the cognitive processes prompted by many utterances just because the information has not been communicated intentionally. On the other hand, relational network theory provides the elements to demonstrate that the hearer actually makes pertinent inferences about the cognitive system of the speaker whom he attributes a lie. For example, in 1989, during his campaign for the presidency, the former President of Argentina, Carlos Saúl Menem used to conclude his speeches with the following words: (7) ¡Síganme! ¡No los voy a defraudar! [Follow me! I won’t disappoint you!] When hearing him, many of us believed that he was lying. Relational network theory allows us to describe and explain such an ordinary phenomenon in linguistic interaction. Figure 8 proposes two relational networks, which account for two different meanings: (a) represents the meanings that Menem intended to evoke in his audience, while (b) describes the interpretation of those who thought that Menem was lying. Concretely, Menem says the he won’t disappoint the people, but many hearers interpret that Menem will disappoint the people. In Figure 8b there is a blocking element coming from a semantic nection: It inhibits the connection between the lexemic nection for no and the semantic nection for NO. It can be said that the concept MENEM blocks the positive interpretation of the whole message; by depicting that the connection between no and NO is blocked, it is represented that this hearer has the representation that can be paraphrased as “Menem will disappoint the Argentine people”
Figure 8. Production and interpretation of an example of “misleading utterance”
AUTHOR’S COPY | AUTORENEXEMPLAR
AUTHOR’S COPY | AUTORENEXEMPLAR
Relevance theory and unintended transmission of information 29 (MENEM FUTURE DISAPPOINT PEOPLE). It is also apparent in this example that relational networks establish a clear distinction between the speaker’s and the hearer’s neurocognitive systems. Intentional and mutually manifest transmissions of information have been the main object of Relevance Theory. However, a theory of language use should also account for unintended meanings and misleading utterances, because they reveal fundamental information about linguistic and cognitive processes. Thanks to Neurocognitive Linguistics and its relational network theory, Pragmatics could widen its perspective in its own benefit by considering all kinds of verbal productions. As a general balance for this section, we could emphasize the following remarks: a. There are many types of utterances whose analysis is highly problematic for Relevance Theory: Unintended puns and Freudian slips, unintentional ambiguities, unintentional contradictions, misleding utterances. b. All the cases mentioned before have a common property: They are actually interpreted as cases of unintended transmission of information. In fact, the hearer makes inferences prompted by those utterances. c. The cognitive process by means of which the hearers actually interpret these cases (Freudian slips, unintended contradictions, misleading utterances, etc.) cannot be accounted for in terms of most core notions of Relevance Theory, such as Communicative Principle of Relevance, informative intention, and communicative intention. d. Some trends in linguistics and philosophy of language have been led to a fallacy. It has been (explicitly or implicitly) concluded that only the intentional transmission of information is worthy of study when considering language use.
5. Conclusions: Toward a neurocognitive approach to language use 5.1. General remarks I will summarize the line of argument developed in the third section of this paper: (i)
Relevance Theory cannot explain the nature of the actual cognitive system needed to interpret unintended transmission of information. (ii) Unintended transmission of information reveals many aspects about the structure of the actual cognitive system underlying linguistic abilities. (iii) Relational network theory accounts for unintended transmission of information. Concretely, it accounts for the neurocognitive system,
AUTHOR’S COPY | AUTORENEXEMPLAR
AUTHOR’S COPY | AUTORENEXEMPLAR
30 José María Gil which enables the individual to produce and interpret such unintended information. Now, the analyses developed within this general line of argument allow us to propose the following conclusions. 1. There are two fundamental Gricean ideas that have been highly influential: (i) The speaker’s meaning is a mutually manifest intention, (ii) the hearer identifies that intention is guided by rational communicative norms. Both of them promoted the analysis of communication as a complex inferential process, but blocked the possibility of considering important samples of linguistic behavior, such as unintended puns. According to Sperber and Wilson (2005: 356) the main subject matters of pragmatics are threefold: (a) implicatures (implicitly communicated propositions), (b) presuppositions, and (c) illocutionary force. Some other important issues such as politeness and deixis could also be included. But independently from the list to complete, it seems to be no real place for the study of unintended transmission of information in Relevance Theory. 2. The linguistic study of Freudian slips could be interpreted as a first and necessary step for providing empirical support to psychoanalytical research. 3. Relevance Theory is good enough to account for cases of intentional transmission of information, i.e., “communication” in its own terms: cases in which there are meanings intentionally evoked by the communicator and evidently identified by the hearer. However, it fails to account for many important instances of verbal behavior. This phenomenon could be called the “intentional fallacy”—the idea that only intentional transmission of information is worthy for linguistics (and philosophy). According to Sperber (1994), a “truly sophisticated hearer” assumes that the communicator may be neither competent nor benevolent, and just intends to seem so. When following the path of least effort in the computation of cognitive effects, this sophisticated hearer should stop not at the first interpretation that is sufficiently relevant, not at the first interpretation the speaker might have thought would be relevant enough to him, “but at the first interpretation that the speaker might have thought would seem relevant enough to him” (Sperber 1994: 16). However, an actual sophisticated hearer should also consider the interpretations that are not only consistent with the speaker’s expectations, but also with the hearer’s needs. Identifying the speaker’s intention is just a part of the cognitive processes involved in verbal comprehension. 4. One of the basic hypotheses of Relevance Theory is that the occurrence of an expression with a given sense allows the hearer to develop a variety of inferences governed by the Principle of Relevance. This seems to be false,
AUTHOR’S COPY | AUTORENEXEMPLAR
AUTHOR’S COPY | AUTORENEXEMPLAR
Relevance theory and unintended transmission of information 31 or only partially true. The hearer develops many inferences that are not guided by such principle; in some cases, those inferences are incompatible with the intended meaning, for example the case of the journalist who was praising Maradona but produced an unintended pun about cocaine addiction. 5. The Communicative Principle of Relevance is incompatible with the cognitive processes that are actually developed by the hearer when making inferences on the basis of an unintended transmission of informa‑ tion. On the other hand, the Cognitive Principle of Relevance is perfectly compatible with those cognitive activities. However, there is a problem with the Cognitive Principle of Relevance and verbal interaction: If we accept that it is the rationale that governs verbal interaction, there are no (substantial) differences between the interpretation of verbal and non- verbal signals. The inference processes promoted by a slip of the tongue yield representations about the speaker’s cognitive system. The inferences processes promoted by the cloudy sky or the yell of a seagull yield representations about the organization or the world. However, it should be acknowledged that the Communicative Principle of Relevance works when considering intentional transmission of information. The Cognitive Principle of Relevance could be complemented with a theoretical framework capable of describing the structure of linguistic information and its relationships with other cognitive systems. This seems to be something that can be done by relational network theory. In fact, the Cognitive Principle of Relevance could be “accommodated” in the terms of Lamb (2006) by Neurocognitive Linguistics. Lamb makes no distinction between linguistic information and non-linguistic information: “We, human beings are engaged in a continuous process of building nections and interconnecting them, and of attempting to influence our fellow creatures in their nection building” (Lamb 2004c: 438). Therefore, “intention” can be conceived as just one of the many varied and complex aspects of human interaction. 6. Defenders of Relevance Theory could argue that “unintended transmission of information” is not communicative at all. However, such an argumentative maneuver would be rather deceitful, because pragmatics intends to account for language use. In section 3, we claimed that some “unintended contributions” are not only part of language use, but they are also actually interpreted by the linguistic and the semological systems of real individuals. Cognitive-Philosophical pragmatics has overestimated the importance of “intention” and assumed that it is at heart of human verbal production and comprehension. Therefore, unintended contributions to conversation
AUTHOR’S COPY | AUTORENEXEMPLAR
AUTHOR’S COPY | AUTORENEXEMPLAR
32 José María Gil are inconvenient phenomena. But any theory can be rescued from falsification if we decide to ignore enough inconvenient phenomena. 7. If relational networks provide acceptable abstract representations of linguistic and cognitive information, we can accept that meaning is always a matter of use. The distinction between semantic and pragmatic meaning has been one of the most important issues within Cognitive-Philosophical pragmatics. Such a distinction seems to be an illusion; it assumes that there is a core meaning encoded by linguistic expressions and that speakers and hearers need to store this core meaning in order to use it properly in interactions. However, relational network theory helps us to understand that linguistic items such as “words” do not have an inner or intrinsic meaning. “Words” do not encode concepts; they are connected to concepts in our highly complex cognitive system (Lamb 1999: 141). (This fact has been represented in Figures 3, 5, 6, and 7). Consequently, there is no need to distinguish between a “core common semantic meaning” and a “contextual/usable pragmatic” one. 8. Relational network theory seems to be a good perspective to study language use as it aims at accounting for all verbal behavior, including unintended transmission of information. It also aims at accounting for the neurocognitive system, which allows the hearers to understand verbal productions. This theoretical framework goes beyond the restrictive concept of “intention.” It conceives that the human cognitive system is in fact a network allowing simultaneous interpretations, for example those needed to understand an intended pun such as the one of the duck in the bar who says, after having drunk a lot, “Put that on my bill.” Of course, the hearer needs to identify that the word “bill” is connected both to the concepts “check” and “beak.” The activation spreads from the lexical nection “bill” (in the interpreter’s linguistic system) to both concepts (in the interpreter’s semantic system), and also to other cognitive systems, such as vision. Figure 9 (slightly adapted from Lamb 2005: 161) represents roughly that both interpretations are relevant and “simultaneous”, since there are parallel spreads of activation in the network.
Figure 9. The pun, simultaneous parallel activation of alternative interpretations
AUTHOR’S COPY | AUTORENEXEMPLAR
AUTHOR’S COPY | AUTORENEXEMPLAR
Relevance theory and unintended transmission of information 33
This overall process is rather different from that described in Table 1, which included the identification of explicatures, implicated premises, and implicated conclusions (proper implicatures). Sperber and Wilson say that these cognitive sub-tasks are not sequentially ordered: The hearer does not first decode the sentence meaning, then construct an explicature and identify an appropriate context, and then derive a range of implicated conclusions. (Sperber & Wilson 2005: 368) But this statement seems to be (unintentionally) contradictory with the very idea of premises and conclusions. That is to say, an interpreter first infers, or posits, premises, and, later, infers a conclusion. In fact, the involuntary contradiction makes manifest in this passage: Which premises are added will depend on the order in which they can be constructed, via a combination of backward inference from expected conclusions and forward inference from information available in memory. (Sperber & Wilson 2005: 370; italics are mine) Examples like the intended pun of the duck are useful to demonstrate roughly that utterance comprehension does not seem to be sequentially ordered. However, it must be so sequentially ordered in terms of Relevance Theory, whose “human cognitive system” needs to combine assumptions conveniently stored in order to find the premises and the conclusions, which are consistent with the Principle of Optimal Relevance. For example, the implicated premises must come after explicatures (see Table 1). The relevance-oriented model cannot account for actual human cog nition, because real neurocognitive (or “inferential”) processes do not seem to follow such an ordered and schematic sequence. In addition, the relevance-oriented model requires the manipulation of symbols, the storing of assumptions, and a computing device capable of manipulating symbols and assumptions, i.e., abstract entities that are not real constituents of the human brain. Relevance-oriented pragmatics needs to accept the existence of those entities because it also assumes explicitly that language is a “closed formal system” and that generative/Chomskyan grammars are the best models of human languages. In Sperber and Wilson’s words, generative grammars successfully account for the semantic representation of a sentence as “a sort of common core of meaning shared by every utterance of it” (Sperber and Wilson 1995: 9). Relational network theory, on the contrary, explicitly aims at satisfying the requirements of operational, developmental, and neurological plausibility. First of all, a network with parallel processing can explain the human cognitive ability to instantaneously access the meanings of linguistic expressions in a conversation; there is no need of search-and-comparison
AUTHOR’S COPY | AUTORENEXEMPLAR
AUTHOR’S COPY | AUTORENEXEMPLAR
34 José María Gil operations, which would be needed by a symbol-based system. Secondly, the requirement of neurological plausibility can be satisfied by the hypo thesis that abstract nections are neurologically implemented as cortical columns (Lamb 1999, 2002, 2005). 9. Sperber and Wilson recognize many times (for example in the epigraph of this paper) that pragmatics should not be concerned only with the representations individually intended by the speaker. The greater the range of alternatives, the weaker the implicatures, and the more responsibility the hearer has to take for the particular choices he makes. Much of human communication is weak in this sense, a fact that a pragmatic theory should explain rather than idealise away. (Sperber and Wilson 2005: 370) Thus, pragmatic theory will explain poetic effects, i.e., the peculiar cognitive effects of the utterances that achieve most of their relevance through a wide array of weak implicatures: Those weak implicatures correspond mainly to the hearer’s responsibility. This proposal suggests that there are many representations that are not intended by the speaker but inferred by the hearer on the basis of the speaker’s verbal behavior. This is consistent with the analysis of relational network theory. However, the proposal seems to contradict the hypotheses about infor mative intention, communicative intention, optimal relevance, etc. If the in terpreter infers “weak implicatures,” which do not deal with the speaker’s intention, then the comprehension process is not guided by the Communicative Principle of Relevance, which states that every act of overt communication conveys a presumption of its own optimal relevance. In such cases, the hearer follows pathways that are different from those required by optimal relevance. In paragraph 3 of these conclusions, it has been said that those pathways could be accounted for by the Cognitive Principle of Relevance, but, according to Sperber and Wilson, this Principle is not about communication. In conclusion, Sperber and Wilson are correct when saying that a pragmatic theory should explain rather than idealize away the non-individually intended representations transmitted by an utterance. However, the core hypotheses of relevance-oriented pragmatics are incompatible with such a profitable end. To sum up: a. If the Communicative Principle of Relevance governs verbal comprehension, there are not weak implicatures (meanings evoked independently from the speaker’s intention), and, b. If there are weak implicatures, the Communicative Principle of Relevance does not govern verbal comprehension.
AUTHOR’S COPY | AUTORENEXEMPLAR
AUTHOR’S COPY | AUTORENEXEMPLAR
Relevance theory and unintended transmission of information 35 10. Thanks to the first evaluation made by the reviewers of Intercultural Pragmatics, I noticed that it was fundamental to emphasize a salient issue: At least part of what the speaker says can be what s/he means (Maybe, part of what the speaker says is always what s/he means). It has also been admitted that, although a slip of the tongue or an unintended pun are samples of non-ostensive behavior, they occur during ostensive behavior. In fact, slips of the tongue or unintended puns exist together with ostensive-inferential communication. Within this context, it should be also emphasized that unintended transmission of information (promoted for example by a slip of the tongue) does not exclude ostensiveinferential communication, and vice versa. In fact, it is being suggested that, probably, the two processes are always together. The problem is that Theory of Relevance accounts for fundamental aspects of ostensive-inferential communication, but it faces serious problems when vaguer effects of ostensive behavior need to be described or explained. It should also be admitted that at least part of what the speaker says is information unintentionally transmitted. Finally, it has not been suggested here that there is a bedrock of stable meaning over and above the speaker’s intention, or even over and above information unintentionally transmitted. In fact, within the neurocognitive approach, it is assumed that the linguistic system changes at any time. It can be expected that the semantic system (which is part both of the linguistic and the wide semological system of the individual) changes every time it is used, that is to say, permanently. 5.2. The argument of the limitation of relevance theory The following list of statements aims at summarizing one of the main arguments in this paper: a. Pragmatic theory should also explain the non-intended meanings evoked in an utterance but not individually intended by the speaker. b. There are several cases of unintended transmission of information, such as unintended puns, unintentional ambiguities, misleading utterances, etc. c. Those utterances in which there is some “unintended transmission of information” do have sense, and they are common phenomena in everyday language use. Consequently, they should be studied by the science of language. d. Those utterances reveal valuable information about the linguistic and semological systems of the speaker. e. The interpretation of those utterances reveals valuable information about the linguistic and semological systems of the hearer.
AUTHOR’S COPY | AUTORENEXEMPLAR
AUTHOR’S COPY | AUTORENEXEMPLAR
36 José María Gil f. When interpreting meanings that have not been intentionally evoked by the speaker, the hearer identifies representations that are (absolutely) independent from the speaker’s intention. g. If the meanings identified by the hearer/interpreter are independent from the speaker’s intention, they have not been derived on the basis of the Communicative Principle of Relevance. h. Therefore, the Communicative Principle of Relevance does not explain fundamental cases that “a pragmatic theory should explain rather of idealize away.” 5.3. From cognitive-philosophical pragmatics to neurocognitive linguistics Thanks to the pioneering work of John Austin, philosophers realized that there was a “Descriptive Fallacy”: To understand the meaning of a sentence was not simply to understand its truth condition. Because of this erroneous belief, philosophers had overlooked many other kinds of utterances and meanings. After the labor developed by Austin, Grice and many other important authors, philosophers, and linguists turned to pay attention to non-descriptive utterances and several aspects of meaning, which go beyond truth value. Currently, we realize that the study of language use may crash into another serious obstacle—we could call it the “Intentional Fallacy.” If linguistics was concerned mainly with intentional transmission of information, it would discard crucial cases of verbal behavior revealing fundamental information about human cognition. Within the general framework of Cognitive-Philosophical Pragmatics, the Theory of Relevance overlapped “intention” and “communication.” Therefore, it has been mainly interested in the information intentionally transmitted by speakers and writers. Thanks to relational network theory, it is possible to suggest that verbal interaction is much more complex (and interesting) than that. Correspondence address:
[email protected] References Arundale, Robert. 2008. Against (Gricean) intentions at the heart of human interaction. Intercultural Pragmatics 5(2). 229–258. Asher, Nicholas and Alex Lascarides. 2003. Logics of Conversation. Cambridge: Cambridge University Press. Atlas, Jay. 2005. Logic, Meaning, and Conversation: Semantical Underdeterminacy, Implicature, and Their Interface. Oxford: Oxford University Press.
AUTHOR’S COPY | AUTORENEXEMPLAR
AUTHOR’S COPY | AUTORENEXEMPLAR
Relevance theory and unintended transmission of information 37 Attardo, Salvatore. 1994. Linguistic theories of humor. Berlin & New York: Mouton de Gruyter. Attardo, Salvatore. 2001. Humorous texts: A semantic and pragmatic analysis. Berlin & New York: Mouton de Gruyter. Attardo, Salvatore. 2006. Cognitive Linguistics and Humor. Humor 19(3). 341–362. Bierwisch, Manfred. 1980. Semantic structure and illocutionary force. In John R. Searle, Ferenc Kiefer, & Manfred Bierwisch (eds.), Speech act theory and pragmatics, 1–37. Dordrecht: Reidel. Blutner, Reinhard and Henk Zeevat. 2003. Optimality Theory and Pragmatics. London: Palgrave. Bou-Franch, Patricia. 2002. Misunderstandings and unofficial knowledge in institutional discourse. In D. Walton & D. Scheu (eds.), Culture and Power: Ac(unofficially)knowledging Cultural Studies in Spain, 323–341. Bern: Peter Lang. Carston, Robyn. 1998. Informativeness, relevance and scalar implicature. In Robyn Carston & Seiji Uchida (eds), Relevance Theory: Applications and Implications, 179–236. Amsterdam: John Benjamins. Carston, Robyn. 2002. Thoughts and Utterances: The Pragmatics of Explicit Communication. Oxford: Blackwell. Cherchi, Marcello. 2000. Investigation of the Brain under Non-pathological Conditions, Section 3. Challenges in the Investigation of Cerebral Function: Neuroanatomical Substrates of Language Processing. LINCOM Europa, 2000. 19–28. Danziger, Eve. 2006. The thought that counts: Interactional consequences of variation in cultural theories of meaning. In Nick Enfield & Stephen Levinson (eds.), Roots of Human Sociality: Culture, Cognition and Interaction, 259–278. Oxford: Berg. Davis, Steven (ed.). 1991. Pragmatics: A Reader. Oxford: Oxford University Press. Davis, Wayne. 1998. Implicature: Intention, Convention, and Principle in the Failure of Gricean Theory. Cambridge: Cambridge University Press. Davis, Wayne. 2007. How normative is implicature? Journal of Pragmatics 39. 1655–1672. Davis, Wayne. 2008. Replies to Green, Szabó, Jeshion, and Siebel. Philosophical Studies 137. 427–445. Dell, Gary S. 1979. Slips of the Mind. LACUS Forum IV, 69–74. Dell, Gary S. and P. A. Reich. 1977. A model of slips of the tongue. LACUS Forum III, 448–455. Dell, Gary S. and Peter A. Reich. 1980a. Slips of the tongue: The facts and a stratificational model. In James Copeland and Philip Davis (eds.), Papers in cognitive-stratificational linguistics: Rice University studies 66(2). 19–34. Dell, Gary S. and Peter A. Reich (1980b). Toward a unified model of slips of the tongue. In V. A. Fromkin (ed.), Errors in linguistic performance: Slips of the tongue, ear, pen, and hand. San Diego: Academic Press, 273–286. Duranti, Alessandro. 2006. The social ontology of intentions. Discourse Studies 8. 31–40. Fitch, W. Tecumseh, Marc Hauser, and Noam Chomsky. 2002. The evolution of the language faculty: Clarifications and implications, Cognition 97. 179–210. Gibbs, Raymond. 1994. The Poetics of Mind: Figurative Thought, Language and Understanding. Cambridge: Cambridge University Press. Gibbs, Raymond. 1999. Intentions in the Experience of Meaning. Cambridge: Cambridge University Press. Gibbs, Raymond. 2001. Intentions as emergent products of social interactions. In Bertram Malle, Louis Moses, & Dare Baldwin (eds.), Intentions and Intentionality, 105–122. Cambridge (MA): MIT Press. Glucksberg, Sam. 2001. Understanding Figurative Language. Oxford: Oxford University Press. Green, Mitchell. 2007. Self-Expression. Oxford: Oxford University Press. Green, Mitchell. 2008. Expression, indication, and showing what’s within. Philosophical Studies 137. 389–398.
AUTHOR’S COPY | AUTORENEXEMPLAR
AUTHOR’S COPY | AUTORENEXEMPLAR
38 José María Gil Grice, H. Paul. 1957. Meaning. Philosophical Review 66. 377–388. Grice, H. Paul. 1967. Logic and conversation. In D. J. Levitin (ed.), Foundations of cognitive psychology: Core readings, 719–732. Cambridge (MA): MIT Press. Grice, H. Paul. 1982. Meaning revisited. In Neil Smith (ed.), Mutual Knowledge, 223–243. London: Academic Press. Grice, H. Paul. 1989. Studies in the Way of Words. Cambridge (MA): Harvard University Press. Hauser, Marc, Noam Chomsky, and W. Tecumseh Fitch. 2002. The language faculty: What is it, who has it, and how did it evolve? Science 298. 1569–1579. Hjelmslev, Louis. 1943. Prolegomena to a theory of language. Madison: University of Wisconsin Press, 1961. Horn, Laurence. 2000. From IF to IFF: Conditional perfection as pragmatic strengthening. Journal of Pragmatics 32. 289–326. Horn, Laurence. 2004. Implicature. In Horn, Laurence & Gregory Ward. 2004. The Handbook of Pragmatics, 3–28. Oxford: Blackwell. Horn, Laurence. 2005. The Border Wars: A neo-Gricean perspective. In Ken Turner & Klaus von Heusinger (eds.), Where Semantics Meets Pragmatics. Amsterdam: Elsevier. Horn, Laurence. 2007. Neo-Gricean pragmatics: A Manichaean manifesto. In N. Burton-Roberts (ed.), Pragmatics, 158–183. Basingstoke: Palgrave. Horn, Laurence and Gregory Ward. 2004. The Handbook of Pragmatics. Oxford: Blackwell. Hubel, David and Torsten N. Wiesel. 1962. Receptive fields, binocular interaction and functional architecture in the cat’s visual cortex. Journal of Physiology 160. 106–154. Hubel, David and Torsten N. Wiesel. 1968. Receptive fields and functional architecture of monkey striate cortex. Journal of Physiology 195. 215–243. Hubel, David and Torsten N. Wiesel. 1968. Functional architecture of macaque monkey cortex. Proceedings of the Royal Society of London 198. 1–559. Ivaskó, L. and Eniko Németh T. 2002. Types and reasons of communicative failures: A relevance theoretical approach. Modern Filológiai Közlemenyek 4. 31–43. Jaszczolt, Katarzyna M. 2005. Default Semantics. Foundations of a Compositional Theory of Acts of Communication. Oxford: Oxford University Press. Jaszczolt, Katarzyna M. 2006. Meaning merger: Pragmatic inference, defaults, and compositionality. Intercultural Pragmatics 3(2). 195–212. Kadmon, Nirit. 2001. Formal Pragmatics: Semantics, Pragmatics, Presupposition and Focus. Blackwell, Oxford. Kasher, Asa. 1986. Pragmatics and Chomsky’s research program. In Asa Kasher (ed.), The Chomskyan turn, 122–149. Cambridge (MA) & Oxford: Blackwell. Kasher, Asa. 1991. On the pragmatic modules: A lecture. Journal of Pragmatics 16. 381–397. Kasher, Asa (ed.). 1998. Pragmatics: Critical Concepts. London: Routledge. Keysar, Boaz. 2007. Communication and miscommunication: The role of egocentric processes. Intercultural Pragmatics 4(1). 71–84. Lamb, Sydney M. 1974. Language as a network of relationships. In Lamb, Sydney. 2004a. Language and reality, 133–175. London & New York: Continuum. Lamb, Sydney M. 1999. Pathways of the Brain: The Neurocognitive Basis of Language. Amsterdam: John Benjamins. Lamb, Sydney M. 2000. Language as real biological system. In Lamb, Sydney. 2004a. Language and reality, 227–253. London & New York: Continuum. Lamb, Sydney M. 2002. On the perception of speech. In Lamb, Sydney. 2004a. Language and reality, 352–357. London & New York: Continuum. Lamb, Sydney M. 2004a. Language and reality. London & New York: Continuum. Lamb, Sydney M. 2004b. Dimensions of the Territory of Neurolinguistics. In Lamb, Sydney. 2004a. Language and reality, 318–323. London & New York: Continuum.
AUTHOR’S COPY | AUTORENEXEMPLAR
AUTHOR’S COPY | AUTORENEXEMPLAR
Relevance theory and unintended transmission of information 39 Lamb, Sydney M. 2004c. Semiotics and Human Knowledge. In Lamb, Sydney. 2004a. Language and reality, 425–443. London & New York: Continuum. Lamb, Sydney M. 2005. Language and Brain: When experiments are unfeasible, you have to think harder. Linguistics and the Human Sciences 1. 151–178. Lamb, Sydney M. 2006. Being Realistic, Being Scientific. In S. J. Hwang, W. J. Sullivan, & A. R. Lommel (eds.), LACUS Forum 32: Networks, Houston. 201–209. Levinson, Stephen. 2000. Presumptive Meanings: The Theory of Generalized Conversational Implicature. Cambridge (MA): MIT Press. Levinson, Stephen. 2006a. Cognition at the heart of human interaction. Discourse Studies 8. 85– 93. Levinson, Stephen. 2006b. On the human ‘interaction engine.’ In: N. Enfield and S. Levinson (eds.), Roots of Human Sociality: Culture, Cognition and Interaction, 39–69. Oxford: Berg. Lewis, David. 1979. Scorekeeping in a language game. Reprinted in David Lewis 1983, Philosophical Papers 1. 233–249. New York: Oxford University Press. Mirecki, Pawel. 2004. Communication ‘failure’: A definition and classification within the Theory of Relevance. In E. Mioduszewska. (ed.), Relevance Studies in Poland 1. 67–87. Warsaw: The Institute of English Studies, University of Warsaw. Mirecki, Pawel. 2005. “Misunderstanding: A starting point for successful communication, A view from the relevance-theoretic perspective. In E. Mioduszewska. & M. Grzegorzewska (eds.), Relevance Studies in Poland 2. 45–52. Warsaw: The Institute of English Studies, University of Warsaw. Mountcastle, Vernon B. 1997. The columnar organization of the neocortex. Brain 120. 701– 722. Mountcastle, Vernon B. 1998. Perceptual Neuroscience: The Cerebral Cortex. Cambridge (MA): Harvard University Press. Neale, Stephen. 2004. This, that, and the other. In Anne Bezuidenhout & Marga Reimer (eds.), Descriptions and Beyond: An Interdisciplinary Collection of Essays on Definite and Indefinite Descriptions and Other Related Phenomena. Oxford: Oxford University Press. Németh T., Enikő. 2008. Verbal information transmission without communicative intention. Intercultural Pragmatics 5(2). 53–176. Németh T., Enikő. 2004. The principles of communicative language use. Acta Linguistica Hungarica 51. 379–418. Nida, Eugene. 1990. The role of rhetoric in verbal communication. Language and communication 10. 37–46. Noveck, Ira, Maryse Bianco, and Alain Castry. 2001. The costs and benefits of metaphor. Metaphor and Symbol 16. 109–121. Paradis, Michel. 1998. The other side of language: Pragmatic competence. Journal of Neuro linguistics 11. 1–10. Recanati, François. 1987. Meaning and Force. Cambridge: Cambridge University Press. Recanati, François. 1989. The pragmatics of what is said. Mind & Language 4. 295–329. Recanati, François. 1998. Pragmatics. In Edward Craig (ed.), Routledge Encyclopaedia of Philosophy 7. 620–633. Recanati, François. 2004a. Literal Meaning. Cambridge: Cambridge University Press. Recanati, François. 2004b. Semantics and pragmatics. In Horn & Ward 2004. 442–462. Reich, Peter. 1985. Unintended puns, LACUS Forum 11. 314–322. Richland, Justin. 2006. The multiple calculi of meaning. Discourse and Society 17. 65–97. Sacks, Harvey, Emanuel Schegloff, and Gail Jefferson. 1974. A simplest systematics for the organization of turn-taking for conversation. Language 50. 696–735. Saul, Jennifer. 2001. Critical studies: Wayne A. Davis, Conversational Implicature: Intention and Convention in the Failure of Gricean Theory. Nous 35. 630–641.
AUTHOR’S COPY | AUTORENEXEMPLAR
AUTHOR’S COPY | AUTORENEXEMPLAR
40 José María Gil Sperber, Dan. 1994. Understanding Verbal Understanding. In: J. Khalfa (ed.), What is Intelligence? 179–198. Cambridge: Cambridge University Press. Sperber, Dan and Ira Noveck (eds). 2004. Experimental Pragmatics. London: Palgrave. Sperber, Dan and Deirdre Wilson. 1995. Relevance: Communication and Cognition. Oxford: Blackwell. Sperber, Dan and Deirdre Wilson. 1998. The mapping between the mental and the public lexicon. In Peter Carruthers & Jill Boucher (eds.), Language and Thought: Interdisciplinary Themes 184–200. Cambridge: Cambridge University Press. Sperber, Dan and Deirdre Wilson. 2005. Pragmatics. UCL Working Papers in Linguistics 17. 353– 388. Stalnaker, Robert. 1999. Context and Content: Essays on Intentionality in Speech and Thought. Oxford: Oxford University Press. Szabó, Zoltan (ed.). 2005. Semantics versus Pragmatics. Oxford: Oxford University Press. Thompson, Robert J. 2008. Grades of meaning. Synthese 161. 283–308. Verschueren, Jef. 1999. Understanding Pragmatics. London: Hodder Arnold. Wilson, Deirdre and Dan Sperber. 2002. Truthfulness and relevance. Mind 111. 583–632. Yus, Francisco. 1998. The ‘what-do-you-mean syndrome’: A taxonomy of misunderstandings in Harold Pinter’s plays. Estudios Ingleses de la Universidad Complutense 6. 81–100. Yus, Francisco. 1999a. Towards a pragmatic taxonomy of misunderstandings. Revista Canaria de Estudios Ingleses 38. 218–239. Yus, Francisco. 1999b. Misunderstandings and explicit/implicit communication. Pragmatics 9(4). 487–517.
AUTHOR’S COPY | AUTORENEXEMPLAR
