Understood variably, meaning is one of the most exalted concepts in human thought, writing, and discourse. Human life itself is discussed in terms of its meaning. Philosophy--one of the highest forms of intellectual enquiry--is defined as an enquiry into the nature and meaning of life and the universe. We question or try to understand the meaning of individuals' actions, government policies, scientific data or theories, philosophical or everyday concepts, a happy or a tragic incidence, or anything that is encountered in life. But it is the meaning of meaning in language that has drawn the greatest intellectual resources from philosophers, linguists, psychologists, poets, literary critics, thinkers in general, and philologists who preceded modern linguists. Speech-language pathologists have both a theoretical and applied interest in meaning because it permeates almost all of their clinical work.
Speech-language pathologists (SLPs) have successfully used applied behavioral techniques in their clinical work. Such applied behavioral techniques as modeling, prompting, shaping, positive reinforcement, differential reinforcement, extinction, time-out, response cost, corrective feedback, among others, are the evidence-based treatment procedures the SLPs use in modifying disorders of speech, language, fluency, voice, and swallowing (Hegde, 1998). Behavioral treatment procedures have been extensively researched and their effects experimentally documented in remediating communication disorders in both children and adults (Duffy, 2005; Hegde, 1998; Hegde & Maul, 2006; Pena-Brooks & Hegde, 2007; Rosenbek, LaPointe, & Wertz, 1989). Although the behavioral treatment techniques used in speech and language training are based on Skinner's experimenta1 analysis of behavior (Skinner, 1953) and the resulting operant conditioning techniques, academic training of SLPs do not seem to include the behavioral (operant) view of verbal behavior to any significant extent (See Table 1). Consequently, SLPs' understanding of speech and language as empirical phenomena on the one hand and of speech and language as targets of treatment on the other are conceptually and methodologically inconsistent. Experimentally validated behavioral treatment procedures are typically grafted on to conceptually inconsistent linguistic, mentalistic, and rationalistic theories that are incapable of experimental verification.
Perhaps it is believed that the behavioral approach may be good for treatment, but not for understanding what language is and how it is learned. Child language treatment research does not support this belief, however. Consistent with Skinner's assertion that linguistic (structural) categories do not correspond to functional response classes, some treatment research studies have shown that grammatical categories and distinctions prove themselves invalid when experimental manipulations are done to treat children's language disorders. Even such basic grammatical-structural distinctions as verbal auxiliary copula and subject noun-object noun do not hold good under experimental manipulations inherent to treatment (Hegde, 1980; Hegde, & McConn, 1978; Hegde, Noll, & Pecora, 1979; McReynolds & Engmann, 1974). Furthermore, it is possible to entertain a conceptually consistent model of verbal behavior (language) and its teaching, but this possibility is realized only with an appreciation of the view that language is behavior. Contrary to what mentalistic linguists argue, the behavioral ana lysis offers a sophisticated and natural science-based analysis of all aspects of language, including what is believed to be the linguist's monopoly--grammar (Skinner, 1957). In this paper, I address not grammar, but meaning, which linguists believe is the stuff the language is made of and lies beyond behavioral explanation. To understand the behavioral analysis of meaning, it is essential to understand the Skinnerian analysis of verbal behavior. To facilitate such an understanding, essential terms of Skinner's Verbal Behavior are defined in Table 1.
A Brief Overview of Skinner's Verbal Behavior
Skinner thought that the term language does not help in a natural science account of what people say, listen to, read, and write. Therefore, he chose the term verbal behavior, which he defined as "behavior reinforced through the mediation of other persons" (Skinner, 1957, p. 2). This definition captures the essence of verbal behavior by including a social community in shaping and maintaining verbal behaviors; the definition is comprehensive in that it includes talking, listening, reading, and writing. Such linguistic structures as phonemes, syllables, words, phrases, and sentences are not analyzed in terms of their causes. Therefore, he proposed a set of new terms that refer to functional response classes that are analyzed in terms of causes and effects. Verbal behavior is classified into verbal operants. A verbal operant is a class of verbal behavior, and a verbal response is an exemplar of a particular verbal operant. For example, a mand is a verbal operant; it includes a variety of verbal responses, all of which are caused by a state of deprivation (motivation), supplemented by certain discriminative stimuli. All kinds of requests, commands, and demands belong to the verbal operant mand. Causes that produce verbal operants as a class and verbal behaviors as specific exemplars are typically described as controlling variables or antecedents. Hunger is a controlling variable for the verbal response, "May I have a hamburger please!" A certain state of the physical environment is the cause or the controlling variable of "It is a beautiful sunset!"--a verbal tact. Printed stimuli are the controlling variables of reading that Skinner called textuals. Please see Table 1 for additional examples.
People do not verbally react to all that they see; responses are evoked by only certain parts of the physical world. People verbally react to stimuli in the presence of which similar responses in the past have been socially reinforced. Following reinforcement, such stimuli are more likely to evoke a verbal response than other stimuli that have not been an occasion for reinforced responses. Stimuli in the presence of which a response is more likely because of previous reinforcement are called discriminative stimuli. Discriminative stimuli may also be called antecedents, and may at least be parts of the controlling variables of verbal responses.
Verbal behaviors are multiply caused. A woman may say "Hi!" to a man or avert eye contact and say nothing. The woman's greeting is determined by the sight of the man as well as the history of their past encounters. We will return to the multiple causation of verbal behavior in understanding meaning from a behavioral standpoint.
Meaning of What?
Meaning in language is different from meaning of verbal behavior. Meaning of verbal behavior can be understood in terms of natural science; but not the meaning in language. Language, defined as a system of symbols or codes that exists independent of human actions does not require a functional (cause-effect) analysis. A structural analysis of printed samples of a language into phonemes, syllables, words, and sentence types will do. It is only when language is thought of as a class of behavior that a need for a functional analysis would emerge. And to think of it as a class of behavior, one needs to abandon the concept of language because it carries too much of a formal (non-empirical), rationalist, cognitive, and mentalistic burden.
An everyday attempt to find meanings of words is to consult a dictionary; but meanings given in the dictionary are meanings of a static language system, not that of dynamic verbal responses. Dictionaries define meanings of words in terms of other words; but they do not define the controlling relationship of words as utterances that are dated empirical events. Meaning, whether of words or sentences, lies in an abstract controlling relationship between empirical events, as explored further in later sections.
A Brief History of Meaning
Meaning has been defined, described, or classified in numerous ways (Mandell & McCabe, 1997; Ogden & Richards, 1923/1989; Smith, 1997; Skinner, 1957; Russell, 1921/2005; 1940/1962 ). Ogden and Richard's 1923 publication lists 16 meanings of meaning (p. 187). Smith's (1997) more recent review of the literature suggests that the conceptual diversity of meaning is by no means reduced in subsequent years. To the contrary, new views of meaning regularly emerge in linguistics and psychology (Allwood & Gardenfors, 1999; Hurford, 2007; Kearns, 2000; McCabe & Mandell, 1997; Smith, 1997; von Heusinger & Turner, 2006). It is not productive to consider all the different concepts of meaning. But it is useful to consider some of the major categories into which the varied definitions may be grouped, although unfortunately, such groupings themselves have varied across investigators. Excluding the behavioral view for the moment, the varied descriptions of meanings in linguistics and philosophy may be classified into four major categories: the bucket theory, the mental (cognitive) entity theory, the referent theory, and the contextual theory.
First, meaning is in the language. Meaning is the content, and words are structures that hold it. Formal linguistic theory may have encouraged this bucket theory of meaning. This is the most commonly described theory in books and articles on language development and linguistics. Bloom and Lahey's 1978 description of language in terms of form-content-use is a prime example of an artificial distinction between the structure and content of language. This view is still popular among speech-language pathologists. Being least analytical, this is perhaps the most commonplace and trite of the different views of meaning.
Second, meaning is a mental or cognitive entity. This entity was more often mental in the past, but it may be, in more recent writings, rephrased as cognitive structures, processes, or mechanical processors (Allwood & Gardenfors, 1999; Smith, 1997). Presumed mechanical units or modules in the head that "process" information are especially popular among cognitivists. For example, a theoretical model on how children learn to read, proposes that children have a semantic processor that stores word meanings and receives excitations from another mechanical devise, the context processor (Adams, 1990). Gardenfors (1999), a cognitivist, is clear about where meaning is and what the slogan should be: "The prime slogan for cognitive semantics is: Meanings are in the head" (p. 21). Mechanical processors are often assigned a home in the head, although some mental entities remain homeless. In any case, all cognitive processors are within-the-skin (mental) activities. When a speaker and listener engage in conversation, certain mental or cognitive entities of the speaker are transferred to, or evoked in, the listener. In this theory, language is the vehicle to transfer meaning from one person to the other. Utterances launch the mental entity of the speaker. Thus airborne, the mental entities land in the mind of the speaker, who then understands the meaning of the utterance heard. This view of meaning is least amenable to a scientific analysis because in an effort to solve one problem (the definition of meaning), it creates a host of more difficult ones. The presumed mental or cognitive entities are harder to observe or explain than the concept of meaning they are meant to explain.
Third, meaning is in the referents of language. This view of meaning, sometimes also described as the denotational meaning (Kearns, 2000), takes into consideration the things and events language is thought to make references to. To justify it, the theory, albeit in its simplest form, points to the things named (Smith, 1997). The meaning of apple is the apple named; the meaning of John is the person so named. Unfortunately, this is another commonplace theory of meaning which easily breaks down when words and utterances that do not refer to physical objects and events are presented for explanation. For example, what is the meaning of such utterances as but, however, nonetheless, if, whereas, in the spirit of, to say the least, and so forth? Language is not limited to nouns, verbs, adverbs, and adjectives that may point to things, or point to images, as some mentalists presume. We will find out later that even for the class of utterances that apparently "stand for" things and events, the referent theory of meaning is not a scientific explanation.
Fourth, meaning is in the context of communication. This theory asserts that the context of language production largely determines its meaning (Smith, 1997). This view is an improvement over the mentalist and referent theories because it takes into consideration the relation between the social context and verbal behavior; the theory avoids mentalistic or cognitive speculations. But context neither always specifies, nor is it always necessary to understand, the meaning of utterances. When someone quotes another person without giving any context, the quote may still "mean" something, though what is meant may be contested by parties involved. The term context is too general, vague to be defined precisely, and too fluid to offer a good explanation. Contexts could be both environmental and intraverbal (see Table 1); neither may be clearly specified. It is difficult to specify the essential parameters of contexts even in specific instances of verbal behavior.
The behavioral scientists' interest in meaning extends beyond its definition. A theory of meaning does not merely define it; such a theory explains it in terms of natural science. Granted, mentalists and cognitivists believe that they, too, offer scientific explanations. But explanations not subjected to the methods of verifications are not scientific. The four kinds of explanations described so far cannot be subjected to scientific verification; their validity is simply taken for granted. Those purported explanations are speculative surrogates for behaviors that can be observed and experimentally manipulated (Skinner, 1977). Crane (2005), in the Oxford Companion to Philosophy, states that "The concept of meaning is every bit as problematic as the concept of mind, and for related reasons. Anyone who conceives of science as objective, and of objectivity as requiring the study of phenomena ... that exist and have their character independently of human thought, will face a problem with the scientific study of meaning" (p. 1, as Retrieved on September 13, 2007).
Mentalistic or cognitive explanations violate a simple rule that scientists follow: do not offer explanations that create additional problems of explanation. As Russell (1926/1986) recommended, it is always good to avoid new problems when trying to solve old ones.
Behavioral Threads in Philosophy of Language
A comprehensive natural science account of verbal behavior was not offered until the publication of Skinner's Verbal Behavior (1957). Available linguistic thought on meaning was (and still is) decidedly mentalistic. In philosophy, however, there has been influential nonlinguistic and nonmentalistic thought on meaning and the nature of language. Some of this thought in philosophy preceded Skinner's analysis of verbal behavior, others were influenced by his analysis. Interestingly, more than linguists, influential philosophers of the early and late 20th century were aligned with a behavioral view of language and meaning. This is not especially surprising, because some of those philosophers were also exponents of the philosophy of natural science who eschewed analysis of an observable phenomenon at another, unobserved, level. More than the linguists, philosophers of natural science were likely to see the value of analyzing language and human behavior from the standpoint of causes and effects.
Several notable philosophers of science, language, and human behavior were unimpressed by the mentalistic linguistic theories of their time. An early provocative view of language that eschewed linguistic mentalistic notions of meaning was expressed by C. K. Ogden and I. A. Richards (1923/1989), both literary critics and philosophers of language. Their critical book, The Meaning of Meaning, rejected the notion that mental imagery is essential to understand meaning. They did not deny the existence of images, but asserted that they do nothing to illuminate meaning in language. They further asserted that most philosophers of language have tried to explain meaning "by inventing a peculiar stuff, an intrinsic property, ... or by inventing a special unanalysable relation" (1989, p. 185). Ogden and Richards were among the first to recognize a need for cause-effect relations in understanding language in general, and meaning in particular. They stated that '"I am thinking of A' is the same thing as to say 'My thought is caused by A ...'" (Ogden & Richards, 1989, p. 55, italics added). They further asserted that "We only know for certain what is said when we know why it is said, though we must not include motives in the 'why'" (Ogden & Richards, 1923/1989, p. 94). They described mentalistic and cognitive notions as "Peopling of the universe with spurious entities ..." (1923/1989, p. 94).
Russell and Other Philosophers on Meaning
Bertrand Russell was perhaps the most distinguished philosopher of science who rejected mentalism of all sorts. An influential contributor to the philosophy of natural science, physics, logic, mathematics, sociopolitical thought, and psychology, Russell was influenced by Watson's (1913, 1924) methodological behaviorism, but improved considerably upon Watson's formulations of language (Wood, 1986)). Russell (1921, 1940) had made his own incisive and nonmentalistic analysis of language and meaning. His review of Ogden and Richard's The Meaning of Meaning, reprinted in 1986 in the Journal of the Experimental Analysis of Behavior, is as refreshing today as it was when it was originally published in 1926. In his review of the Ogden and Richards book (1923/1986), Russell (1986 reprint, p. 109-110) summarized 8 truisms that he held about words:
* "Words are social;" this may be an early concept of verbal behavior as a class of social behavior.
* "Words are bodily movements;" words belong to classes (an early hint of Skinner's verbal response classes), and each instance of a word is a bodily movement.
* "Words are means of producing effects on others;" another precursor to Skinner's analysis of the way listeners are affected by speakers' verbal operants.
* "Words, like other bodily movements, are caused by stimuli;" stimuli may be external or a state of the body (a tooth ache, for example); the concept of "thoughts" is evoked when the speaker cannot locate a physical stimulus in his or her body; one of the early recognitions of stimuli and responses, albeit in the manner of Watson.
* "Heard words are stimuli;" a viewpoint that includes, like Skinner's does, speaker and listener in an analysis of verbal interactions.
* "It is not of the essence of words to express 'ideas';" an analysis of stimuli and responses is sufficient; there is no need to postulate intermediary ideas--a radically different view of language.
* "The distinction between the emotional and the logical use of words is illusionary;" effects the words create are more important in understanding meaning, rather than some subsidiary covariables.
* "In the individual, the heard language is earlier than spoken language;" an infant's behavior is affected by verbal stimuli before the infant can affect others' behavior through his or her own verbal behavior.
These are remarkable statements, written in 1926 by a philosopher of natural science. Russell had a clear concept of stimuli, responses, and the effects verbal responses had on others. The instant he considered language a form of social behavior, he moved away from mentalistic images and cognitive entities. Russell stated that "The relation of a word to its meaning is of the nature of a causal law governing our use of the word and our actions when we hear it used. There is no more reason why a person who uses a word correctly should be able to tell what it means than there is why a planet which is moving correctly should know the Kepler's laws" (2005, P. 118; original publication, 1921). Again, similar to Skinner's view, Russell suggested that an explanation of "knowing a language" need not evoke mentalistic or cognitive entities; it only needs to evoke an analysis of causal relationships (Russell, 1940/1962). Knowing the meaning of words, Russell asserted (1940/1962), is a matter of saying those words in appropriate situations and "acting appropriately when they are heard" (1962, p. 24). Therefore, in a causal analysis of meaning, grammatic distinctions or knowledge of word meanings are both unimportant and misleading.
In his later theory of language, the influential Ludwig Wittgenstein (1958--a posthumous publication of an earlier work) formulated language in terms of social use and the effects language production have on others--another nonlinguistic and nonmentalistic view of language. Wittgenstein rejected many commonly held linguistic assumptions, including the referent theory of meaning, the idea that meaning is a mental process or entity, and the suggestion that meaning lies in the intentions of speakers (Day, 1969; Finch, 1977; Pole, 1958; Wittgenstein, 1958). Among other influential philosophers who rejected mentalism and cognitivism, Quine (1960) is noteworthy. He suggested that meaning as a mental entity and reference as a theory of meaning are obscure intermediaries that may well be abandoned. His major work, Word and Object, was influenced by Skinner's Verbal Behavior.
One well-known American linguist who rejected the mentalistic theory of language as the expression of feelings, ideas, images, thoughts, and so forth was Bloomfield (1933). He formulated language in terms of stimuli and responses based on the Pavlovian conditioned reflex model. Bloomfield analyzed acts of speech which have practical events preceding them and similar events following them. He considered meaning determined by the events that precede and follow an act of speech. Although it seems to have influenced Skinner (Passos & Matos, 2007), Bloomfiled's analysis of language was later overshadowed by the Chomskyan transformational generative grammar of the 1950s and subsequent generative syntactic and cognitive semantic theories within linguistics (Passos & Matos, 2007; Allwood & Gardenfors, 1999; Kearns, 2000).
Analysis of meaning in nonlinguistic and nonmentalistic philosophy was noteworthy in its rejection of traditional views of meaning in favor of a cause-effect analysis. Although such philosophical analyses, especially those of Russell, are closer to Skinner's than most linguists', there is something missing in the philosophical analyses described so far. What is missing is Skinner's concept of behavioral contingency. Russell for example, recognized the importance of stimuli, the need to keep the analysis at the level of observation, and the effects the verbal behaviors have on listeners. For Russell as well as Ogden and Richards, stimuli and environmental events are causes of utterances (verbal operants). Their analysis of effects of verbal behaviors typically stops at the listener; missing is a thorough analysis of the listeners' actions in shaping verbal behaviors. While the speaker affects the listener's behavior, the listener, too, affects the speaker's behavior. It is indeed the listener's contingent behavior that has the effect of shaping the verbal operants of the speaker. Furthermore, their analysis does not tell why only certain stimuli, not all stimuli one might encounter, are causes of verbal behavior. An answer to why and how only certain aspects of the physical world come to control verbal behaviors in an individual requires the concept of reinforcement contingencies that have the power of ontogenetic behavioral selection. Causes and effects should consider the interlocking verbal operant paradigm (Winokur, 1976) in which the effects the listeners produce for the speaker should be analyzed. In essence, a more complete analysis of the contingencies of reinforcement, which include the effect the social community has on the speaker in shaping verbal behaviors, was left for Skinner.
Meaning in Skinner's Early and Later Writings
Skinner's Verbal Behavior was published in 1957, but he had been working on it for many prior years. His early analysis of verbal behavior was concerned with poetry and other forms of literature. Skinner has written interesting papers on literature (e.g., Skinner, 1934/1972; 1939/1972; 1941; 1972) that were a precursor to his more thorough analysis of verbal behavior. This is not surprising because Skinner's early interest was in a literary career.
The concept of meaning explored in Skinner's writings on literature contrasts with his later analysis presented in Verbal Behavior. His pre-1945 writings on poetry and other forms of literature considered meaning a property of words, reference a relation between words and the things referred to, and meaning a potential determinant of verbal behavior (Andery, Micheletto, & Serio, 2005). But in his post-1945 writings, especially in Verbal Behavior, Skinner completely rejected his own earlier formulations of meaning. Along with a more thorough analysis of verbal behavior, he then proposed a new way of talking about meaning in terms of the controlling variables of verbal operants (Skinner, 1957). Although Skinner continued to write on verbal behavior (e.g., 1974, 1986), he never significantly revised his 1957 view of meaning.
With only 4 indexed entries for the term meaning in Verbal Behavior, Skinner did not address the issue extensively, though he made his position clear: Meaning is (a) not properties of words or sentences; (b) not to be found in the heads or minds of speakers; (c) not in the ideas, intentions, feelings, or thoughts the verbal behavior is presumed to convey; (d) not in the referents; (e) not something communicated or expressed by speakers; (f) not something listeners mentally grasp. One might add that meaning is not in the simple context of utterances either. He rejected the possibility that in a natural science account, something like meaning exists at all.
Skinner's Definition of Meaning
Skinner did not believe that a reformulation of mentalistic concepts of meaning in scientifically useful terms will serve any purpose. Such a reformulation will only help perpetuate the rest of the mentalistic baggage. But something that plays such a central role in both lay and academic discourse could not be entirely dismissed; so he stated that:
... Meaning is not a property of verbal behavior as such but of the conditions under which behavior occurs. Technically, meanings are to be found among the independent variables in a functional account, rather than properties of the dependent variable. When someone says that he can see the meaning of a response, he means that he can infer some of the variables of which the response is usually a function. (Skinner, 1957, pp. 13-14)
Skinner's definition of meaning as the relation between verbal operants and their causal (controlling) variables has been discussed by other behavioral scientists (Andrey, Micheletto, & Serio, 2005; Lowenkron, 2004; Moore, 2000, among others). Much of the behavioral analysis of meaning (the entire range of verbal behavior for that matter) is unknown to many SLPs. Most students in speech-language pathology learn about verbal behavior from its critics who do not understand it in the first place and are biased in favor of the linguistic approach. Therefore, an attempt is made in this paper to extend the Skinnerian analysis of verbal behavior in general, and meaning in particular, to a variety of child language phenomena as well as some "abstract" forms of verbal behavior the linguists think create difficulties for the behavioral view.
Extension of Skinner's Analysis of Meaning
To fully understand the implications of Skinner's definition of meaning, it is essential to consider it as the functional (causal) relationship between verbal responses and their controlling variables which include motivational states and environmental events. Because verbal operants may be caused by such bodily states as hunger and pain, it is essential to include motivational states in the analysis of meaning. However, meaning is neither in the motivational states (including "intentions'), nor in the environmental events, nor even in the verbal responses themselves. It is in this sense that meaning is not a mental or objective entity, or a response property. Meaning is a function of a consistent relationship between a certain verbal response or class of verbal responses, their controlling variables, and consequences that follow the behaviors. Verbal responses are treated as meaningful only when such a consistent causal relationship is recognized by the members of the verbal community who serve as audience. If meaning is abstract, it is because it lies in a relation between three sets of events: verbal behaviors, their antecedents, and the consequences. Meanings would be much less abstract if words or sentences simply conveyed them, elusive if they were to be found in the heads and minds of listeners because there are no means to objectively extract them, and vague if they were to be found in the context of utterances unless we can specify the properties of the context that are the antecedents to the response. Fortunately, verbal behavior is particularly potent in revealing this abstract relation.
Verbal Operants Have Multiple Controlling Variables
As noted, each verbal operant may have multiple causes. Even mands, that are determined by bodily states (e.g., hunger and thirst), are not solely determined by such states. A man who is hungry is likely to mand food only in the presence of certain discriminative stimuli: inside a restaurant, sight of food, inside a home where someone could supply food (reinforce the mand), and so forth. A man who walks into his house and announces to his wife that "Bhutto was assassinated yesterday" (a tact) has a conditioning history of talking about Bhutto with his wife, a current stimulus that may include an announcement on the radio, TV news, or a newspaper, and the presence of his wife who is the audience (discriminative stimulus). Speech is more or less likely depending on the occasion; an occasion is a set of discriminative stimuli in the presence of which certain verbal operants have been reinforced in the past. In essence, all verbal operants are determined by a past conditioning history and a present occasion that sets them off.
To give effective response to verbal operants of speakers, the listener needs access to the controlling variables of those operants. In everyday language, to give a correct response, the listener needs to "understand the meaning" of what was said. But should the speaker tact (specify) all of the controlling variables of his or her speech to be "understood?" Probably not. In many cases, the listener already has access to some of the controlling variables, but lacks access to others, perhaps to some critical ones. In such cases, the speaker is not understood fully. As the examples in the section entitled Further Explications and Extensions make it clear, speakers tact one or more of the multiple controlling variables of their speech for the benefit of the listener.
Verbal Operants Can Point to Some or All of Their Own Controlling Variables
That certain verbal operants reveal their own controlling variables may be one of the unique properties of verbal behavior. While nonverbal behaviors do not explain why they are being produced, verbal behaviors may. For instance, when a man gets up from his chair and begins to walk across the room, the walking itself does not announce why it is taking place, although its cause may be inferred from past experience. The woman who is watching the walking may infer that the man is going to the bathroom. But it is also possible that the woman does not "understand" why the man is walking, and may ask "Where are you going?" Therefore, meaning as defined is not unique to verbal behavior. Nonverbal behaviors, too, can be meaningful (the cause inferred or verbally stated) or meaningless (the cause unknown and verbally unspecified) depending on whether the controlling variables are clear or obscure.
A verbal operant (an utterance) that reveals its own controlling variable has two parts: a primary verbal operant, and the secondary verbal operant that tacts (reveals, describes) the controlling variable of the primary verbal operant. In everyday terms, an utterance may consist of a part that says something to the listener and a part that says why something is being said. What is being said is primary, and the accompanying explanation of it is secondary; they are secondary because without the primary verbal operant, there is no need for the secondary verbal operant. The part that says why something is being said reveals the controlling variable of what is being said. In more technical terms, secondary verbal operants tact the controlling variables of primary verbal operants and both may be parts of a single utterance. This is done entirely for the benefit of the listener. Speakers are typically in touch with the controlling variables of their speech, although this does not necessarily mean that speakers can describe the causes of their verbal behavior. People do not always understand the reason why they said or did something. Speakers speak because the controlling variables have produced their effects on them.
When speakers speak under specified stimulus conditions that are a part of the controlling relationship, they may specify why they just said something. Among the multiple controlling variables, some may be obvious to the listener; such variables are not specified. For instance, under certain stimulus conditions that are obvious to the listener (e.g., a news paper article or a discussion of personal investment being discussed) speaker may say "The stock market will improve this year because past years of presidential elections have seen a surge in stock prices," contains a primary verbal operant (in roman script) and a secondary verbal operant (in italics). The secondary verbal operant tells the listener the main reason why the primary verbal operant is currently being emitted: the speaker would not have predicted improved stock prices if such improvement was not seen in previous years of presidential elections. If speakers did not specify certain aspects of controlling variables of their primary verbal operants, listeners may force them to say why they are saying or why they said something. The answer to such demand will specify the needed access to the critical feature of the controlling variables of what was said, or in some cases, what is about to be said. Note that the utterance in the example does not tact any other controlling variable (e.g., past history of similar talk) that might also be the part of the reinforcement contingency.
The assertion that some verbal responses tact critical aspects of controlling variables of other responses does not mean that meanings of some responses are conveyed through, or carried by, other responses. Meaning is always a function of the controlling variables of verbal operants; only some, but not all, (secondary) verbal operants carry the burden of specifying one or more controlling variables of the primary verbal operants so the listener can "understand" them.
That one part of a verbal response can tact the controlling variables of another part has not escaped the attention of philosophers. In recognizing certain semantic paradoxes, philosophers have acknowledged that a sentence can talk about itself (Semantic paradoxes, in Oxford Dictionary of Philosophy, 1996; Oxford Reference Online).
Autoclitics Tact Certain Controlling Variables of Primary Verbal Operants
Verbal responses which tact the (or at least some of the) antecedents of other verbal responses are called autoclitics in Skinner's (1957) analysis; they are also secondary verbal operants. All secondary verbal operants are autoclitic s. Not all autoclitics tact the controlling variables of primary verbal operants, however, but those that do are always autoclitics. Autoclitics include not only the topographically sequenced order found in utterances and the traditional grammatical morphemes, but also other types of responses that linguists may not include under grammar. For instance, such utterances as "I said ..." or "I was about to say ..." are secondary autoclitics in Skinner's analysis, but linguists do not consider them as parts of grammar. Utterances of that kind are descriptive autoclitics which simply describe a previous or an imminent response, but do not reveal the relationship between the primary verbal operants that accompany them and their antecedents (causes). Many other autoclitics point to that relationship. For instance, a speaker might say "I said it is going to be a good year because many economists have predicted it." In this verbal operant, the italicized term is an autoclitic (a secondary verbal operant), which also describes the controlling variable of the non-italicized part of the primary verbal operant. Additional examples throughout the paper will further illustrate this point.
Brown (1973) suggested that grammatic al morphemes modulate meaning. In behavioral terms, most grammatical morphemes are autoclitics that help specify some aspects of the controlling variables of primary operants. Take for instance:
* "I see two cups." The plural morpheme s, an autoclitic, tacts the numeric property of the controlling stimulus.
* "The boy is walking." The auxiliary is, another autoclitic, tacts the temporal relation to the utterance and its controlling variable; the present progressiveing essentially does the same.
* "The ball is on the table." The autoclitic on, a preposition, tacts the spatial relation between the ball and the table, which together, form the controlling variables of the primary verbal operant.
* "It is Mommy's hat." The autoclitic possessive 's tacts the relation between the person and the object, which are the controlling variables of the verbal operant.
An analysis of this kind can be extended to most grammatical morphemes to show that their function is to tact some aspects of the controlling variables of the primary verbal operant for the benefit of the listener. In this kind of behavioral analysis, the relation between grammar and meaning--a subject of intense speculation with mentalistic notions in semantics--is clear. Grammatical elements, by tacting aspects of the controlling variables of verbal operants, help listeners gain access to the total set of those variables.
Mands Need Not Tact Their Controlling Variables
There is one class of verbal operants that need not typically specify their controlling variables. They are mands. To use a cliche, mands wear their causes on their sleeves. The controlling variables of mands are typically transparent to the listener. When someone says "May I have a glass of water, please?," the listener normally does not ask "Why?" The cause of that request (mand) is clear: thirst, a physiological state of the speaker's body. The mand is a discriminative stimulus for the listener, who reinforces it because of the past conditioning history. Also because of that history, the listener has had in the past, and currently has access to, the controlling variable s of such mands. In common terms, the listener "knows" what it means to demand a glass of water and proceeds to reinforce the mand by offering it.
In the behavioral ana lysis, the statement that the "listener knows" only means that the listener has access to the controlling variables of the particular mand (or any other class of verbal operants). In other words, mands of the kind have been parts of the listener's learning history; no cognitive or mental entities are invoked. There may be reasons to question the validity of mands (or any other verbal operants), however, as we will see in a later section on truth and falsehood of verbal operants.
Further Explications and Extensions
To fully understand the concept that meaning is a function of the controlling relation between verbal operants, their antecedents, and consequences, we can examine several examples of everyday conversation, rather than manufactured artificial sentences linguists use to illustrate their assumptions. Although the structural units of language are irrelevant in a causal analysis, the behavioral view of meaning in terms of a controlling relationship can be illustrated at all structural levels--word, sentence, and conversational speech as the example s that follow illustrate. Note that in each case, an utterance of any length or a conversational episode becomes meaningful only when the listener gains access to some of the multiple controlling variables:
1. Some word productions may not tact the controlling variables. A two-year-old toddler boy says "truck," but the parents do not understand its meaning, even though they "know what the word means." Thus, "knowing the meaning of a word" in the traditional analysis, and understanding the meaning (the independent variable) of the same word production when it is a specific verbal response, are different phenomena. Parents, in this example, need access to the causes of the child's utterance. When the parents ask, "What do you mean, Johnnie?," they are not asking for the meaning of the word the child just produced; they are seeking access to the controlling variable of that response. The child might then show the picture of a truck in a story book he has been looking at, to which the parents may respond by saying, "Oh Yes! I see it now!" As this example illustrates, speakers (especially young children) may provide access to the critical or current controlling variables of their verbal operants by nonverbal behavior as well.
2. In some conversational exchanges, the controlling variables systematically unfold. The following illustration involves two persons, A and B.
(A) "How come you didn't show up last evening?"
(B) "What do you mean?"
(A) "You didn't come to the meeting last night."
(B) "What meeting?"
(A) "The faculty meeting."
(B) "Oh, I didn't know there was one."
In this example, B does not "understand" the meaning of A's questions or statements until the end, illustrating clearly that words and sentences do not carry meaning to the speaker, and that meaning is not a property of verbal responses. B does not understand A precisely because he has had no access to the causal relation between certain events and A's verbal responses ("How come you didn't show up last night?" and "You didn't come to the meeting last night"). Finally, B understands A when it is established that there was a faculty meeting the previous night, B was not there, and that these two events stand in causal relation to A's current (at least the initial) verbal response. Linguists may say that as the exchange continues, the "meaning becomes more and more clear" or that "the speaker's ideas are mapped onto the listener;" but none of these explain the listener's initial problem. What really becomes clearer to the listener (as well as to you, the reader of that transcribed exchange) is the relation between A's current responses and their antecedents (the faculty meeting and B's absence).
3. Speakers may tact the controlling variables of responses made in the past. The child's utterance, "I said it is bad because yesterday you said it is bad" illustrates this case. The part of the response beginning with because tacts the controlling variable of the most recent response; the recent response itself is a delayed echoic, and that is what the first part specified.
4. Typically, speakers are conditioned to tact some of the controlling variables of their verbal operants. A speaker says, "The weatherman predicts it is going to snow tonight." If the speaker simply said that "It is going to snow tonight," the listener may have asked "Why do you say so?" Or, "How do you know?" Any time a listener questions a speaker about why something is being said, the listener in essence is demanding access to the independent variables (causes) of the speaker's verbal responses. Such listener reactions enter into the contingency that shapes the speaker's verbal operants that include autoclitic tacts of their own verbal operants. One might say that the speaker in our example preempted the listener's question by adding an autoclitic ("The weatherman predicts"--the secondary verbal operant) to the primary verbal operant. Other similar everyday examples include: (a) "I saw on the evening TV news that Benazir Butto was assassinated;" (b) "Tom told me that he will be in Hawaii this week;" (c) "You were the one who has been saying that we should buy this car;" (d) Jane called me earlier to say that a faculty meeting will be held today." In each of these examples, the italicized portions are secondary autoclitics that tact (specify) some of the controlling variables of the primary verbal operants. Obviously, each of the illustrated verbal operants has additional controlling variables (e.g., a past history of interactions and reinforcement contingencies) without which it would not be emitted in the first place.
5. In a group situation, some may have access to the controlling variables of a speaker's speech, but others may not. A speaker facing several listeners gathered in a room, asks, "Shall we go?" Some of the listeners then say, "Yes" while others ask, "Go where?" Apparently, the meaning is clear to some, unclear to others. To say that the listeners who said, "Yes" got the speaker's idea, but those who asked, "Go where?" did not get the idea is no explanation; it simply restates the problem: the listeners either did or did not understand the utterance (or that they did or did not get the idea). Technically, those who agreed with the speaker have had prior access to the event that controlled the response, "Shall we go?" (in every day language, "they knew what the speaker was talking about."). The rest have not had that access, and therefore, what was not clear to them was the relation between "Shall we go?" and its antecedent controlling event.
6. Children may generalize verbal operants without tacting their controlling variables. The author's son Manu, at age 2:1 said, "Egg and cheese, dad" while looking at a toy car. When asked "Where?" he pointed at yellow and white spots of glue on the car's chassis. The original explanation of such generalized or "overextended" responses, well documented in child language (Clark, 1973; Hoff, 2005), was that the child had learned only a few of the multiple semantic features of the words that are overextended (Clark, 1973, 1974). Overgeneralizations, however, are a matter of stimulus generalization and do not require mentalistic or cognitive-semantic explanations. To understand the meaning of overgeneralized responses, the listeners only need to see the stimulus properties that enter into the controlling relationships of such responses. In our example, the meaning is in the controlling relation between the child's response and the current and the original stimulus properties (the white and yellow spots of glue and the actual fried egg with cheese). It would normally be assumed that the resemblance between egg/ cheese on the one hand and yellow/white spots of glue on the other makes the child's response "meaningful." Resemblance in this and similar cases is a function of shared stimulus properties that control a response, and meaning is a function of listener's access to that controlling relationship.
7. Children learn to tact progressively more of the controlling variables of over-generalized verbal operants. Although psycholinguists have suggested that overgeneralizations decline after age two, there is observational evidence that children do not simply stop producing them, but instead learn to tact more of their controlling variables. For example, the author's son, at age 2:3, began adding the word like before every overextension. Responses such as like egg and cheese, and like a hammer while hitting an object with a spoon became more frequent after 2:00. Still later, around 3:00, what would have been overextensions were qualified by looks like (looks like moon in relation to a patch of yellowish color in a magazine) and looks like with further qualifications (e.g., looks like Bert, but not exactly) in relation to a muppet on the TV show, Sesame Street. Such additions ("looks like" or "not exactly") tact the stimulus properties of accompanying generalized responses; the added tacts help refine the listener response ("who understand that the child is responding on the basis of stimulus similarity"). Such overgeneralizations do not need mentalistic or cognitive explanations; they are simply generalized responses. That they are socially undesirable is beside the point in a causal analysis.
8. Children are under overly-selected stimulus control when they produce "overly -restricted" or "underextended" words. Another child language phenomenon, which is the opposite of overgeneralization, has been called over-restriction or underextension (Bloom, 1973; Nelson, 1974; Bowerman, 1978). A typical example is provided by Bloom (1973) whose daughter Allison at 9 months used the word car only when she could watch cars moving down the street from her living room window. She did not use the word for pictures of cars or actual cars in other situations. The linguistic explanation of this phenomenon is that the child has learned only a few of the multiple meanings of the word. This is no explanation but a rephrase of the problem to be explained. Furthermore, it is not the utterance that is overly restricted, as psycholinguists imply, but it is the controlling stimulus complex that is overly restricted (selected). Instead of all features of a complex discriminative stimulus controlling a verbal operant, only one or two features of that stimulus may initially control the response--resulting in overly discriminated responses. The overly restricted controlling stimulus is a strong discriminative stimulus, possibly because of initial social reinforcement for the response given in its presence. With differential reinforcement, the child's responses generalize so that multiple features or a total stimulus complex will begin to control the emission of verbal operants.
9. The terms "meaning," "meaningful," and "meaningless" are also verbal operants most speakers routinely emit. As verbal operants, such utterances may have nothing to do with their own controlling variables; they may mostly be about certain consequences the speaker faces in his or her life. As a verbal operant, "meaningful" tacts reinforcing consequences and "meaninglessness" tacts the lack thereof. A man who says, "I don't find that statement meaningful" is not directly tacting (commenting on) the controlling variables of the statement. The man (speaker) is essentially saying that (a) he would not make the same statement under similar conditions; (b) the statement holds no personal reinforcing consequences; and (c) the statement may be false, in the sense described later. The woman who says that someone else's verbal operants are "meaningful" is essentially saying that (a) her behavior may be positively affected by them; (b) she would emit similar verbal operant under similar stimulus conditions; and (c) she may judge the statement as true, in the sense described later. Similarly, the verbal operant "What does it mean to me?," is more a query of potential reinforcers from whatever being tacted, rather than a pursuit of something lofty. "My existence is meaningless"--another verbal operant, a deeply philosophical one at that, tacts the deficient reinforcing conditions of one's life. A philosopher's question, "What is the meaning of life?" seeks to analyze the controlling variables and reinforcing consequences of life in general, certain specific living conditions, or specific actions. In essence, analysis of "meaning," "meaningful," and "meaningless," when they are verbal operants (as someone's utterances), must be included in a complete analysis of meaning. Meaning as a controlling relation of verbal operants in general is different from the controlling variables of "meaning" as specific verbal operants.
Semantic Categories and Response Classes
In linguistics and psycholinguistics, semantic categories are various features of meaning that children learn. Each word is thought to have multiple semantic features. In this sense, meaning is a property of words. It is proposed that children's task is to map the different semantic features onto their cognitive domains. Various linguistic theories proposed that early two- and three-word utterances reflect semantic categories or features (Schlesinger, 1971, 1974; Bloom, Lightbown, & Hood, 1975; Bowerman, 1973, 1978; Leonard, 1976; Wells, 1974). Some of the most commonly cited semantic categories include action, agent, existence, location, negation, possession, recurrence, attribution, nomination, instrument, denial, rejection, and object. These categories are defined in terms of what the child knows or recognizes. For example, one of the definitions of agent states that it is "... the recognition that an animate entity initiated an activity" (Leonard, 1976, p. 6l); instrument is defined as the specification of an "... inanimate object that was used in an action to affect another object" (Bloom et al., 1975, p.13). These semantic categories are typically said to "emerge" in the child's speech. Such an emergence is often attributed to Piaget's stages of cognitive development.
Cognitive psycholinguists hold that words are not labels for objects but rather tags for concepts. When an utterance reflects a certain semantic category, the child is said to possess a corresponding semantic concept (meaning). For example, utterances of the kind the doll is on the table are said to imply that the child is able to recognize a spatial relationship between two objects, and the child is also said to possess the concept of location. Utterances of the kind he was hit by the ball are taken to mean that the child has recognized that an inanimate item was causally involved in an activity, and therefore, the child has the concept of instrument. The child who said the grass was cut is assumed to show recognition that an inanimate object was receiving the force of an action; the child is also presumed to possess the concept of object. Thus, the child is hypothesized to acquire various semantic categories and learn rules for using words belonging to those categories. These rules specify the "conditions that must obtain before the word can be used appropriately or correctly" (Bowerman, 1978, p. 113).
Psycholinguists often ignore the fact that semantic categories, notions, or relations are the result of a formal analysis. It is not evident that children actually classify objects or events in terms of instrument, attribution, location, and so forth. For example, there is no evidence to assert that the child who said the grass was cut was thinking in terms of an inanimate object receiving the force of an action; the child certainly could not verbalize it. It is doubtful even if most adults can. The fact that the child can utter such a sentence is no evidence that he or she possesses a concept of instrument. As pointed out by Brown (1973, p. l46), semantic relations (or categories) "... are abstract taxonomies applied to child utterances." But many psycholinguists drive those taxonomies back into the heads of children. Furthermore, the assertion that words are tags for concepts, not objects, may help avoid the problems inherent to the reference theory of meaning, but the assertion transforms meaning into an even more obscure cognitive entity.
Semantic categories illustrate cognitive theorists' typical attempt at classifying observable behaviors with no reference to their independent variables. A reduction in the classification of dependent variables (utterances, etc.) can be achieved when their independent variables are analyzed, because it is quite possible that a variety of structurally different utterances are controlled by the same or similar independent variables (Skinner, 1957; Winokur, 1976). As Skinner's analysis has shown, such a cause-effect analysis will yield far fewer functional response classes than a structural analysis.
What is Meaningless?
Theoretically, if meaning is carried by the words as the bucket theory of meaning asserts, there shall be no meaningless utterances as long as they include the words the listeners have "understood" in the past. Only unfamiliar words and foreign language will be meaningless to listeners. But interestingly, not understanding the meaning of a verbal response is not the same as judging that response meaningless; listeners do not necessarily judge the utterances they do not understand as meaningless. The listeners often assume that the utterances are, or may be, meaningful, but that they just "don't get it." Behaviorally, this means that the listeners have not encountered contingencies of reinforcement that would have shaped that response; they have not been in a situation where they had produced the same or similar utterance. That is why the listeners demand access to at least some of the controlling variables of utterances they do not understand ("seek clarification" as commonly put). When the speaker allows that access by "explaining why something he or she said," the listener can imagine conditions in which they themselves or others might produce the same or similar utterance. Utterances are the same or similar if they produced the same or similar effects on the listeners; structural or topographic similarity linguists describe is not implied.
The preceding underscores another important fact of verbal operants. While the listener who is denied access to the controlling variables of verbal operants may not understand the meaning of utterances and demand access, those who do obtain the access may, at times, judge the utterance meaningless, or more likely, false. Therefore, access to controlling variables of verbal operants only makes it possible for the listener to understand why a particular verbal behavior was emitted. It has nothing to do with "grasping" the meaning of the verbal operant. Understanding the causes of a verbal behavior may not produce the effect the speaker had hoped-for. We will return to the behavioral account of true and false verbal operant in a later section. For now, it suffices to reiterate that in most typical verbal exchanges, "meaningfulness" is a function of the controlling variables of heard verbal operants, and "meaninglessness" is a function of controlling relationship that is either obscure to the listener or questionable from the standpoint of the listener's past experiences. Such judgments are by no means cognitive entities, however. They are short-hand description of the cumulative effects of the verbal contingencies the listener was exposed to in the past. In essence, the listener has experienced similar contingencies, or knows others who have.
Understanding and Agreeing are Different Phenomena
Understanding the meaning of the verbal operants of a speaker and agreeing with the speaker are not the same, as everyone knows. A listener may understand, but not agree with the speaker. What does this mean? It means that a listener can understand a verbal operant because the controlling variable of that verbal operant was known. But in everyday experience, and in more formal academic discussions, one could judge what is said is nonsense, not because it was not understood, but what was understood was disagreeable. In behavioral language, "I understand what you mean, but I don't agree with you" means that "I know why you said what you said, but I would not say it under similar stimulus conditions." As stated before, understanding or not understanding utterances are a matter of gaining or failing to gain access to their controlling variable s. What follows after one understands a verbal operant is a matter of different kinds of behavioral analyses, not that of meaning in terms of a controlling relation between verbal operants, their antecedents, and potential consequences.
Disagreement is an interesting verbal operant that typically evokes cognitive explanations. It is the difference in thinking, entertaining contradictory ideas, or experiencing cognitive dissonance, and such other mentalistic notions that are said to explain disagreement among people. In the behavioral analysis, two people disagree when the contingencies that generated a statement are not shared by them; the speaker and the listener have been exposed to different conditioning histories relative to the verbal operant in question. It is the most objective explanation in the sense that there is no value judgment as to who is right and who is wrong. This does not mean, however, that verbal operants cannot be behaviorally analyzed in terms of truth and falsehood, validity and invalidity. Such important and often mentalistic terms can be analyzed in terms of natural science.
Apparent meaninglessness can also be defined in relation to the behavioral history of an entire verbal community. In this context, meaningless verbal responses are those for which the verbal community does not have typically experienced controlling events and potential reinforcing consequences. Also, the verbal community in such conditions may not have a verbal operant either ("they don't have a word for it"), but if one is introduced, the verbal community may treat such responses as either meaningless or anomalous, as described later. Nonsense syllables and other meaningless verbal responses can be manufactured, however. Linguists do this routinely, as the famous Chomskyan "sentence" Colorless green ideas sleep furiously. But such meaningless utterances have their own controlling variables and the contingencies involved there can be described. But those variables usually do not figure in the common reinforcement history of the verbal community.
Meaning of Anomaly, Ambiguity, Paraphrase, and Synonyms
The behavioral analysis of meaning is able to account for a number of related verbal phenomena such as anomaly, ambiguity, paraphrase, and synonym, in a conceptually consistent manner (Skinner, 1959, Winokur, 1976). A "sentence" or a verbal response is anomalous when the listeners cannot recognize a set of independent variables for it, even though the response topography is acceptable (i.e., has certain commonly occurring structural properties). "The flying pig..." is anomalous precisely because the verbal community has not experienced an event that could possibly control the natural production of that response and perhaps could not be reinforced if produced. Nonsense syllables, meaningless verbal responses, and anomalous sentences are all the same in this respect. They differ only in terms of response topography.
Ambiguity, on the other hand, results when a particular verbal response is typically controlled by different discriminative stimuli, and the response itself is nonspecific to its actual antecedent. For instance, the verbal operant "They are eating apples" produced with even stress on all syllables is ambiguous not because of its two different deep structures, but because it does not tact its controlling variables. In addition, an ambiguous response serves as a discriminative stimulus to many verbal and nonverbal responses, and this also creates problems to the listener because the competing responses cannot be checked. Simply put, an ambiguous response can have different, and sometimes contradictory, effects on the listener. In everyday discourse, ambiguity is resolved not by diving into the deep structures, nor by checking the speaker's intention, but by asking questions that evoke answers which tact the antecedents of utterances. Quite often, manufactured sentences that are ambiguous in a linguistic analysis are not so at all when they are verbal operants. When a speaker says "They are eating apples!" it is very likely that the controlling event will be there to see: some edible apples or some people eating apples. Consequently, the response is not ambiguous.
A paraphrase is a sequence of verbal or written responses which are controlled by the same discriminative stimuli as the phrases (responses) they substitute. In other words, the contingencies governing the phrase and its paraphrase are the same or similar with corresponding effects on the listener. A synonym also functions similarly, but in this case the response topography is limited to single words.
An analysis of utterances that listeners judge true, false, honest, dishonest, correct, incorrect, or deceptive, will show that the concept of the controlling variables of verbal operants is capable of accounting for a wide range of verbal behavioral phenomena in terms of natural science. Such phenomena are typically explained by appealing to hypothesized within-the-skin variables, including good or bad, moral or immoral, and ethical or unethical minds.
When is a verbal utterance true? When is it false? In other words, what is the functional analysis of verbal behaviors judged true or false? Are truth and falsehoods related to meaningfulness? Behavioral answers to these questions suggest that truth, falsehood, and meaningfulness are all a function of the controlling variables of verbal statements, or more comprehensively, the actions the verbal behaviors tact and the effects such tacts have on the listeners. Contrasting judgments people make are not cognitive entities or processes in the behavioral analysis. Instead, they are differential effects the verbal behaviors have on listeners; such effects are almost always described in terms of different responses listeners give to the speakers.
We have noted that autoclitics talk about the controlling variables of primary verbal operants and that philosophers, too, have recognized that sentences talk about themselves. But a sentence may tact true or fake controlling variables. Obviously, even false utterances are meaningful in the sense defined: they must have their controlling variables without which the utterances would be impossible. An autoclitic may tell truth or lie about the controlling variables of the primary verbal operants it is a part of. In simper terms, one part of the sentence may tell the truth, or lie about, the other part. A statement is deemed false (reacted to as false) when the true controlling variables of it are concealed but the speaker tacts socially more acceptable or self-serving controlling variables. Speakers who specify socially acceptable causes of their utterance while concealing their true and socially unacceptable causes are likely to receive reinforcement from listeners. In other words:
1. A tacted controlling variable may have been the cause of the verbal behavior (VB); if so, the VB is true (reacted to as such by listeners).
2. A tacted controlling variable may not have been the cause of the VB; if so, the VB is false (reacted to as such by listeners).
Truth and falsehood have provoked immense philosophical and epistemological debate; but in a behavioral analysis, true and false judgments are differentiated listener responses--both verbal and nonverbal. Such differential listener responses are based on the kinds of access they gain to the controlling variables of verbal operants. This means that the verbal utterances themselves are neither false nor true. Truth, like meaning, is not a property of the verbal responses, nor is it a property of the person who tells truth. Truth is not hidden in utterances that a "truthful person" produces. Neither is falsehood hidden in verbal operants of liars; false statements are not those cracked verbal buckets from which truthful juice has somehow leaked-out.
Verbal operants do not convey truth or falsehood to the listeners. For example, President Bush has said that the presence of weapons of mass destruction is the reason why he invaded Iraq. Some people have deemed his statement false (or meaningless), but what exactly is deemed false (or meaningless)? Not the verbal statement per se, because the statement itself is a fact. It is the tacted controlling variable of the act of invasion that some people consider false because they believe that the true controlling variables of invasion are different. Such people also infer that the reinforcement contingencies implied by the president (enhanced national security) are not true; they infer other kinds of reinforcing contingencies for the President's action.
In the case of nonverbal behaviors that are tacted by a speaker (as in the previous example), a verbal operant is true if both the speaker and listener agree on the controlling variables of the nonverbal actions the utterance tacts. A secondary criterion of judging an utterance true is the agreed-upon reinforcing consequences of actions (e.g., "I did it for reasons of national security"). Furthermore, a speaker who says that someone's utterance is true is essentially saying that under similar stimulus conditions, he or she would say the same. Conversely, a speaker who says that someone's utterance is false is essentially saying that under similar stimulus conditions, he or she would not say the same. Implied in both the cases is a past conditioning history.
A false utterance that tacts a fake controlling variable of an action may be further judged honest or dishonest. An honest utterance tacts the true controlling variable of actions ("an honest utterance gives the real reason of actions taken"); concealing its true controlling variables, a dishonest utterance specifies socially more reinforcing (acceptable) reasons for actions. Thus, in this sense, a false and a dishonest utterance are both the same: both are lies in everyday language. But some lies are honest whereas others are dishonest--both are differential listener reactions (judgments).
An honest lie is a verbal operant that correctly (honestly) tacts the controlling variables of an action at the time the verbal operant is emitted, but eventually it turns out that the variables tacted were not the true variables of the action. When the utterance is emitted, it is judged true or correct because at the time the person spoke (or wrote), he or she did believe that the tacted controlling variables were indeed true variables of the action in question. (Speakers do not always have access to the controlling variables of their own verbal operants, as noted before.) Possibly, listeners may not have access to the true controlling variables either. In defense of an honest liar, one might say that "She later realized that her statement was not true." It essentially means that the actual controlling variables of previously incorrectly and tacted controlling variables of an action later came to light.
A dishonest lie is a verbal operant that tacts fake controlling variables of actions, and the speaker had access to the true controlling variables at the time the utterance in question was emitted. "He knew he was telling a lie," a distracter might say.
Listeners of honest or dishonest lies may or may not describe themselves as deceived. Listeners may agree with the speaker that the tacted controlling variables of the speaker's actions were indeed true (though they were false) and act accordingly. But if the listeners later find out that the true controlling variables of actions were different from what the speaker had tacted, they will then describe themselves as having been deceived. Listeners, who from the beginning, rejected the speaker's description of the controlling variables of his or her actions do not describe themselves as deceived. They will not have acted in a way to reinforce the speaker, either. Such listeners are likely to say, "I knew from the beginning that he was lying, so I did not support him."
Only the Meaning of Tacts and Autoclitics May be True or False
As noted, there are two kinds of tacts: nonautoclitic (primary verbal operant) tacts that "talk about" the physical world and its events and autoclitic tacts that point to the controlling variables of those primary verbal operants. Only the autoclitic and nonautoclitic tacts are subjected to true or false judgments relative to meaning. Other classes of verbal operants, including mands, echoics, textuals, and intraverbals are not judged for their truthfulness or falsehood.
Examples given in the previous sections show that autoclitics may tact the wrong controlling variables of primary verbal operants. An additional example may clarify it further. A man said that "I read in the newspaper that it is going to rain today." It contains an autoclitic (in italics), which tacts the controlling variable of the rest of the current statement. That autoclitic, as well as the nonautoclitic tact, it is going to rain today, could be subjected to true/false judgments (differential listener responses). The man who said it may not have read it in the newspaper; he may be telling a lie. Perhaps he did not want to go out for a walk in the evening, but his wife wanted to, so he lied about the rain. The autoclitic I read in the newspaper is false because it tacts the wrong controlling variables of it is going to rain. The true controlling variable of it is going to rain is the aversive consequences of going for a walk and the reinforcing consequences of staying home.
Purely nonautoclitic tacts also may be judged true or false. For example, when a speaker says, "That is a beautiful gardenia!" the listener may either say, "Yes! It is!" or "No, it is a jasmine!" The listener's agreement or disagreement technically strengthens or weakens the tact, but at the moment, may also be described as a response that implies a judgment of truth or falsehood.
The meaning of mands, however, is not typically subjected to a judgment of truth or falsehood. For example, a woman asks, "I am thirsty. Can I have a beer?" Only that autoclitic tact part of it (in italics) can be true or false. That the woman wants a beer--the mand in the verbal operant--is not judged for its truthfulness; the listener will take it for granted. When only the controlling variable of "I am thirsty" is suspect, can that mand be questioned, but what is questioned is not the fact that the woman wants a beer, but that she is thirsty (she is not). The listener may overtly (most likely covertly) say that "She is not really thirsty, but wants to drink a beer anyway." The effect the mand has on the listener (whether the listener gives a beer or not) may influence the probability of that and similar mands being emitted in similar situations in the future, but for the moment, the "reason for the request"--the controlling variable of the mand--is judged false.
Echoics, too, are not judged true or false, though they may be judged correct or incorrect. Provided an echoic reproduces or approximates its own stimulus, the listener may reinforce it as correct verbal behavior. But neither the behavior nor its controlling variable is true or false. Similarly, a textual, which is under the control of printed stimuli, may be judged correct or incorrect, but not as true or false. Intraverbals, which are verbal operants under the control of prior verbal operants, illustrate interlocking chain of verbal behaviors. Technically, they are a matter of chained verbal stimuli, evoking chained verbal operants, not subjected to judgments of truth or falsehood, but may be judged correct or incorrect.
Summary and Conclusions
A natural science-based analysis of behavioral phenomena seeks causes and eschews appeal to entities presumed to exist at unobserved (and unobservable) levels. If meaning is to be explained in terms of natural science, one cannot accept mentalistic and cognitive explanatory inventions. Listeners who react to verbal behaviors of a speaker do not somehow mentally reach out and pull out mental or cognitive entitie s from the head of that speaker. Listeners do not analyze the deep structures of utterances to find the meaning hidden there. Words do not carry meaning, nor do they all refer to things and persons. In a behavioral analysis, what semanticists describe as meaning is an abstract relation between the verbal behaviors and their controlling variables. In a general sense, "understanding an utterance" is the same as gaining access to the controlling variables of speakers' verbal behavior. In simpler terms, "to understand the meaning of an utterance" is to know why a speaker said what he or she did. In essence, the listeners need to know the causes of verbal behaviors to say that they "understand their meaning." The cause-effect relation that characterizes the traditional notion of meaning of verbal behavior can account for varied child and adult verbal behavior, including such abstract phenomena as truth and falsehood of verbal operants. The traditional and exalted concept of meaning, once redefined in cause-effect terms, serves no further purpose.
I would like to thank Raymond Weitzman, Professor Emeritus of Linguistics, California State University-Fresno for his constructive criticism that helped me greatly in revising this paper. I would also like to thank the anonymous reviewer for helpful comments for revision.
Adams, M. J. (1990). Beginning to read: Thinking and learning about print. Cambridge, MA: The MIT Press.
Allwood, J., & Gardenfors, P. (Eds.) (1999). Cognitive semantics: Meaning and cognition. Philadelphia: John Benjamin's Publishing Company.
Andery, M. A., Micheletto, N., & Serio, T. M. (2005). Meaning and verbal behavior in Skinner's work from 1934 to 1957. The Analysis of Verbal Behavior, 21, 163-174.
Bloom, L. (1973). One word at a time: The use of single-word utterances before Syntax. The Hague, Netherlands: Mouton.
Bloom, L., & Lahey, M. (1978). Language development and disorders. New York: John Wiley.
Bloom, L., Lightbown, P., & Hood, L. (1975). Structure and variation in child language. Monographs of the Society fur Research in Child Development 40,(2), Serial No. 160.
Bloomfield, L. (1933). Language. New York: Holt, Rinehart, and Winston.
Bowerman, M. (1973). Early syntactic development. London: Cambridge University Press
Bowerman, M. (1978). Semantic and syntactic development: A review of what, when and how in language acquisition. In R. L. Schiefelbusch (Ed.), Bases of language intervention (pp. 97-190). Baltimore, MD: University Park Press.
Brown, R. (1973). A first language: The early stages. Cambridge, MA: Harvard University Press.
Clark, E. V. (1973). What is in a word? On the child's acquisition of semantics in his first language. In T. E. Moore (Ed.), Cognitive development and the acquisition of language (pp. 65-110). Now York, NY: Academic Press.
Clark, E. (1974). Some aspects of the conceptual basis for first language acquisition. In R. L. Schiefelbusch & L. L. Lloyd (Eds.), Language perspectives: Acquisition, retardation, and intervention (pp.105-128). Baltimore, MD: University Park Press.
Crane, T. (2005). "Meaning." The Oxford companion to philosophy. Oxford Reference Online [http://www.oxfordreference.com]. Oxford, England: Oxford University Press.
Day, W. F. (1969). On certain similarities between the philosophical investigations of Ludwig Wittgenstein and the operationism of B. F. Skinner. Journal of the Experimental Analysis of Behavior, 12, 489-506.
Duffy, J. R. (2005). Motor speech disorders: Substrates, differential diagnosis, and management (2nd ed.). St. Louis, MO: Elsevier-Mosby.
Finch, H. L. (1977). Wittgenstein: The later philosophy. Atlantic Highlands, NJ: Humanities Press.
Gardenfors, P. (1999). So me tenets of cognitive semantics. In J. Allwood & P. Gardenfors, Cognitive semantics: Meaning and cognition (pp.19-36). Philadelphia: John Benjamin's Publishing Company.
Hegde, M. N. (1980). An experimental and clinical analysis of grammatical and behavioral distinctions between verbal auxiliary and copula. Journal of Speech and Hearing Research, 23, 864-867.
Hegde, M. N. (1998). Treatment procedures in communicative disorders (3rd ed.). Austin, TX: Pro-Ed.
Hegde, M. N., & Maul, C. (2006). Language disorders in children: An evidence-based approach to assessment and treatment. Boston, MA: Allyn and Bacon.
Hegde, M . N., & McConn, J. (1978). Language training: Some data on response classes and generalization to an occupational setting. Language training: Some data on response classes and generalization to an occupational setting. Journal of Speech and Hearing Disorders, 46, 353-358.
Hegde, M. N., Noll, N. J., & Pecora, R. (1979). A study of some factors affecting generalization of language training. Journal of Speech Hearing Disorders, 44, 301-320.
Hoff, E. (2005). Language development (3rd ed.). Belmont, CA: Wadsworth/Thomson Learning.
Hurford, J. R. (2007). The origins of meaning. New York, NY: Oxford University Press.
Kearns, K. (2000). Semantics. New York, NY: St. Martin's Press.
Leonard, L. B. (1976). Meaning in child language. New York, NY: Grune and Stratton.
Lowenkron, B. (2004). Meaning: A verbal behavior account. The Analysis of Verbal Behavior, 20, 77-97.
Mandell, C., & McCabe, A. (Eds.) (1997). The problem of meaning: Behavioral and cognitive perspectives. New York, NY: Elsevier Science.
McCabe, A., & Mandell, C. (1997). The many meanings of meaning. In C. Mandell & A. McCabe (Eds.), The problem of meaning: Behavioral and cognitive perspectives (pp. 1-14). New York, NY: Elsevier Science.
McReynolds, L. V., & Engmann, D. L. (1974). An experimental analysis of the relationship of subject and object noun phrases. In L. V. McReynolds (Ed.), In L. V. McReynolds (Ed.), Developing systematic procedures for training children's language. ASHA Monograph No. 18.
Moore, J. (2000). Words are not things. The Analysis of Verbal Behavior, 17, 143-160.
Nelson, K. (1974). Concept, word, and sentences: Interrelations in acquisition and development. Psychological Review, 81, 267--285.
Ogden, C. K., & Richards, I. A. (1923/1989). The Meaning of meaning. New York, NY: Harcourt Brace Jovanovich.
Pasos, M. L. R. F., & Matos, M. A. (2007). The influence of Bloomfield's linguistics on Skinner. The Behavior Analyst, 30, 133-151.
Pena-Brooks, A., & Hegde, M. N. (2007). Assessment and treatment of articulation and phonological disorders in children (2nd ed.). Austin, TX: Pro-Ed.
Pole, D. (1958). The later philosophy of Wittgenstein. London, Athlone Press.
Quine, W. V. (1960). Word and object. Cambridge, MA: MIT Press.
Rosenbek, J. C., LaPointe, L. L., & Wertz, R. T. (1989). Aphasia: A clinical approach. Austin, TX: Pro -Ed.
Russell, B. (1921/2005). The analysis of mind. Mineola, NY: Dover Publications, Inc. [Originally published in 1921 by Allen & Unwin, London & Macmillan, New York.]
Russell, B. (1926/1986). Bertrand Russell's review of the Meaning of Meaning. Dial, 1926, 81, 114-121. Reprinted with an introduction by W. S. Wood in the Journal of the Experimental Analysis of Behavior, 1986, 45, 107113.
Russell, B. (1940/1962). An enquiry into meaning and truth. Baltimore, MD: Penguin Books. [Originally published in 1940 by Allen & Unwin, London.]
Schlesinger, I. M. (1971). Production of utterances and language acquisition. In D. I. Slobin (Ed.), The ontogenesis of grammar (pp . 63-101). New York: Academic Press.
Schlesinger, I. M. (1974). Relational concepts underlying language. In R. L. Schiefelbusch & L. L. Lloyd, (Eds.), Language perspectives: Acquisition, retardation and intervention (pp. 129-151). Baltimore, MD: University Park Press.
Skinner, B. F. (1934/1972). Has Gertrude Stein a secret? In B. F. Skinner, Cumulative record (3rd ed., pp. 359-369). New York, NY: Appleton Century Crofts.
Skinner, B. F. (1939/1972). The alliteration in Shakespeare's sonnets: A study in literary behavior. In B. F. Skinner, Cumulative record (3rd ed., pp. 385-390). New York, NY: Appleton Century Crofts.
Skinner, B. F. (1941). A quantitative estimate of certain types of sound-patterning in poetry. American Journal of Psychology, 54, 64-79.
Skinner, B. F. (1953). Science and human behavior. New York: The Free Press.
Skinner, B. F. (1957). Verbal Behavior. New York, NY: Appleton Century Crofts.
Skinner, B. F. (1972). A lecture on having a poem. In B. F. Skinner, Cumulative record (3rd ed.), (pp. 345-355). New York, NY: Appleton Century Crofts.
Skinner, B. F. (1974). About behaviorism. New York: Knopf.
Skinner, B. F. (1977). Why I am not a cognitive psychologist. Behaviorism, 5, 1-10.
Skinner, B. F. (1986). The evolution of verbal behavior. Journal of the Experimental Analysis of Behavior, 45, 115-122.
Smith, L. D. (1997). Historical and philosophical foundations of the problem of meaning. In C. Mandell & A. McCabe (Eds.), The problem of meaning: Behavioral and cognitive perspectives (pp. 15-80). New York: Elsevier Science.
von Heusinger, K., & Turner, K. (2006). When semantics meets pragmatics. Boston, MA: Elsevier.
Watson, J. W. (1913). Psychology as the behaviorist views it. Psychological Review, 20, 158-177.
Watson, J. W. (1924). Psychology from the standpoint of a behaviorist. Philadelphia: Lippincott.
Wells, G. (1974). Learning to code experience through language. Journal of Child Language, 1, 243-269.
Winokur, S. (1976). A primer of verbal behavior: An operant view. Englewood Cliffs, NJ: Prentice-Hall.
Wittgenstein, L. (1958). Philosophical investigations (G. E. M. Ansocombe, Trans). Oxford, England: Basil Blackwell.
Wood, W. S. (1986). Introductory comments on Bertrand Russell's review of the Meaning of Meaning. Journal of the Experimental Analysis of Behavior, 45, 107-113.
Author Contact Information
M. N. Hegde
1948 Ashcroft Avenue
Clovis, CA 93611
e-mail: email@example.com (e-mail contact is preferred)
Table 1. Essential terms of Skinner's (1957) Verbal Behavior. Skinner's terms Examples Antecedents Operant All non-reflexive and Various environmental Behaviors that act voluntary behaviors events (stimuli) and (operate) on the reinforced and internal states of the environment and sustained by their body (deprivation or generate consequences. aversive stimulation). consequences. Verbal operant A class or a kind of Each kind has a A functional class of verbal behavior; not different antecedent verbal responses technically, but as described in the with similar causes loosely may refer to a subsequent sections of and effects; include particular instance of this table. tacts, mands, verbal behavior--a intraverbals, specific verbal echoics, textuals, response. and autoclitics are verbal operants. Any utterance may be a verbal operant Tacts "Flower!" Parts of the physical (Descriptions, "It is a table." world (environmental comments, etc.) "He is John." events, persons, A primary verbal "It is hopping." objects, etc; external operant. stimuli that cause sensory consequences) Mands "May I have some A physiologic state of (Commands, water please." deprivation or demands, requests, "Shush!" ("Be quiet") aversive stimulation questions, etc.) "What do you mean?" (hunger, thirst, A primary verbal "Your place or my discomfort, pain, operant. place?" sexual deprivation); different states of the body, as against the events outside the skin. Echoics Clinician: "Say, Another person's Imitative book." verbal behavior (as behaviors that Child: "Book." against the physical reproduce their The child's response aspects of the world own stimuli. is an echoic. around or internal A primary verbal states of the body) operant Intraverbals Almost all continuous Immediately preceding Verbal behaviors and connected speech verbal responses of generated by the (and writing). the speaker that immediately prior Recitation of the stimulate additional verbal behaviors of alphabet; proverbs and verbal responses. For the speaker. common sayings; example, in reciting A primary verbal typical phrases (e.g., the alphabet, naming operant. table--chair); history the prior letters are as recited; all the stimuli for the sentence completion following letters; tasks; most academic uttering the name of talk; all narratives one president may with sequence and stimulate the name of temporal order. the next president; saying "table" may stimulate "chair." Autoclitics All grammatical Primary verbal Parts of verbal morphemes plus any behaviors, without operants that tact part of an utterance which there will be no the controlling that tacts the need for the secondary variables of primary controlling variables autoclitics. verbal operants. of primary verbal For instance, the A secondary verbal operants. primary reason why the operant, a special "I see two cups." quantifier two and the kind of tacts; called "The boy is walking." plural s in the first secondary because "I heard it over the example are emitted is their emission radio that the that the speaker is depends on the democratic currently tacting a primary verbal Presidential candidate stimulus comp lex (two operants. is coming to town." cups). The primary tact is the antecedent of two and the morpheme s. Textuals Naming the printed Printed stimuli for (Reading and alphabets reading. writing) Naming the words and Printed stimuli for sentences (oral and copying. silent reading). Oral speech for Writing: copying the dictated writing. letters, writing to Intraverbal control dictation, writing for "spontaneous under intraverbal writing." control ("spontaneous writing" or composition) Skinner's terms Consequences Operant Positive, negative, or aversive Behaviors that act events that follow behaviors and (operate) on the help increase or decrease their environment and future probability. generate consequences. Verbal operant Each kind is shaped and A functional class of maintained by different kinds of verbal responses antecedents as described in the with similar causes subsequent sections of this table. and effects; include tacts, mands, intraverbals, echoics, textuals, and autoclitics are verbal operants. Any utterance may be a verbal operant Tacts Listener's agreement or (Descriptions, disagreement: comments, etc.) "A beautiful flower!" A primary verbal "That's right, it is a table." operant. "No, he is not John!" "Hopping fast!" Reinforcement is typically verbal. Mands Primary reinforcers that listeners (Commands, offer to reduce deprivation or demands, requests, aversive stimulation. questions, etc.) Presentation of food items; A primary verbal removal of aversive stimulation; operant. opportunities for sexual behaviors; reinforced by nonverbal as well as verbal behaviors. Echoics Presence or absence of Imitative reinforcement, depending behaviors that strictly on whether the reproduce their response reproduces the own stimuli. stimulus, approximates it, or A primary verbal moves in the right direction. operant Intraverbals Such simpler intraverbals as A-B- Verbal behaviors C-D or common sayings (table-- generated by the chair) are explicitly socially immediately prior reinforced; more complex verbal behaviors of intraverbals, such as a lecture that the speaker. describes sequenced historical A primary verbal events, reinforcement is more operant. subtle; both social reinforcement and subtle discriminative kinesthetic cues may play a role; delayed social reinforcement that comes after an extended period of speech also may strengthen intraverbals. Autoclitics Social reinforcement typically in Parts of verbal the form of listener agreement or operants that tact discriminated verbal and the controlling nonverbal behavior. variables of primary For instance, the listener may verbal operants. react differently to the second A secondary verbal example depending on whether operant, a special the boy is or was walking. kind of tacts; called Autoclitics are for the benefit of secondary because the listener, not speaker. their emission depends on the primary verbal operants. Textuals Social reinforcement for a (Reading and progressively better writing) correspondence between reading and the printed stimuli ("correct reading."). Social, conditioned generalized, or primary reinforcers (monetary gain) for "spontaneous writing."…