Previous | Contents | Next

CHAPTER II. Bloomfield's "meaningless" science of sounds

The study of language can be conducted without special assumptions so long as we pay no attention to the meaning of what is spoken. (Bloomfield 1933:75)

This statement is definitive of where meaning fit into Bloomfield's linguistics. It didn't. This isn't to imply there is nothing to say about Bloomfield's semantics. Indeed, there is a great deal to be said about where Bloomfield situated meaning, and the process through which a statement such as the one above could come about.

When Bloomfield's contribution to linguistics is summed up very briefly, it is often with the phrase "the discovery of the phoneme". He is credited as being the first to explicitly state that the object of linguistic study is the smallest significant unit of sound and what may be directly built upon that. Linguistics, in the decades immediately preceding Language, was only just emerging as an academic discipline. In the spirit of the empirical times, with a penchant for reductionism and hard-nosed science reaffirmed by Weiss, and spurred on by the great number of world languages (particularly the accessible Amerindian languages) which lacked written descriptions, Bloomfield sought to provide both a reliable methodology and a rigorously defined object of study.

In keeping with what Bloomfield had learned from Weiss's behaviorism, he focused explicitly on what could be observed and stated in precise scientific terms. For linguistics to function as a science it was necessary to eliminate "assumptions" which could not be scientifically verified, reducing language phenomena to its very basic constituents. Recall the story of Jack and Jill in Chapter One, and the "significant" events or stimulus-responses (A and C) versus the "pure" speech act (B).

In principle the student of language is concerned only with the actual speech (B); the study of speakers' situations and hearers' responses (A and C) is equivalent to the sum total of human knowledge. If we had an accurate knowledge of every speaker's situation and of every hearer's response - this would make us little short of omniscient - we could register these two facts as the meaning (A-C) of any given utterance (B), and neatly separate our study from all other domains of knowledge. The fact that speech-utterances themselves often play a part in the situation of a speaker and in the response of a hearer, might complicate things, but this difficulty would not be serious. (Bloomfield 1933:74)

The meaning is here defined as the two facts (A-C). The hunger and salivation of Jill as she sees an apple and the heroism of Jack as he jumps a fence, climbs a tree, and fetches the apple. In this instance Bloomfield (1933) defined the meaning as being quite external to the organisms involved. And, meaning is defined as being quite external to linguistics.

Linguistics on this ideal plane would consist of two main investigations: phonetics, in which we studied the speech event without reference to its meaning, investigating only sound-producing movements of the speaker, the sound-waves, the action of the hearer's ear-drum, and semantics, in which we studied the relation of these features to the features of meaning, showing that a certain type of speech-sound was uttered in certain types of situations and led the hearer to perform certain types of response. (1933:74)

Actually our knowledge of the world in which we live is so imperfect that we can rarely make accurate statements about the meaning of a speech-form. The situations (A) which lead to an utterance, and the hearer's responses (C), include many things that have not been mastered by science. (1933:74-75)

So, what exactly is Bloomfieldian linguistics? Since his scientific and methodological requirements made semantics untenable, there remained only the other of the two main investigations identified: phonetics. This chapter is meant to be an examination of meaning in Language (1933), however to focus exclusively on what little was said about meaning and ignore the corpus that exists around it would be artificial. So it will be useful to see what Bloomfield was trying to do. Though these sections of Language are devoid of "meaning", they are far from meaningless to this thesis.

Bloomfield did not only point his readers in a direction or suggest an approach, he offered a complete methodology which could be carried out without variation from one linguist to the next. The basic item, the atom of the linguist if you will, was the phoneme. Bloomfield defined the phoneme as "a minimum unit of distinctive sound-feature." (1933:79). There were a finite set of phonemes in any given language which could be detailed by a linguist in the field and thus provide the basic building blocks of a particular language. All phrases, all spoken utterances in a language, were composed of individual phonemes strung together serially, and it was the task of the linguist to find minimal pairs - words such as pin and bin which are two distinct words in English showing that English speakers make a distinction between the phonemes /b/ and /p/.

If phonemes functioned as atoms in Bloomfield's linguistics then phonetics can be seen as representing the table of elements. This analogy is of no stretch to anyone who has seen a typical phonetic chart as used by linguists with the various "places of articulation" listed across the top like the periods of elements. Phonetics is the expression of all possible sounds or, more properly, phones in human language. This represents one of the few pieces of information that a linguist does not have to gather on a strictly language-by-language empirical basis, and a pragmatically necessary time-saver and standardization tool for a linguist in the field.

A linguist would transcribe a speech utterance into a series of phones that would be analyzed for distinguishability and collected into classes of phonemes. At each step of analysis there was a prescribed set of "discovery procedures" which would yield data for a new level of analysis. Thus the raw phonetic data, analyzed into phonemes, can be used in the next level of analysis - morphology. The chemistry analogy may be extended to this level by comparing morphology to molecular study. A chemist, having defined what atoms are in a solution, may, through spectrometric analysis, find how these atoms are arranged into "meaningful" molecular units.

A linguist following Bloomfield's program would, through a process no less extravagant than spectrometry, look for meaningful units. He or she would analyze as many strings as possible, searching for repetitions, in different environments, of the smallest possible string. These presumably, were words or word primitives. One additional step is to look for complementary distribution and phonemic similarity between pairs of morphemes to determine if they are actually, essentially the same morpheme. Thus it could be determined that the morphemes "knife" and "knive_" never occur in the same environment and can be considered morpheme alternants. Again, a chemist would recognize that the crystals of ice and gaseous water vapor never occur in the same environment but represent different aspects of the same molecule. In both cases, generalizations about the process involved would be attempted.

The data from the morphological level may be directly fed into a "discovery procedure" for the sentence-syntax level. The morphemes are lined up and, based on regularities in function, a linguist would be able to make certain inferences on clause or phrase structure. Also at this stage is the possible determination of morpheme class based on function - presumably nouns, verbs, adjectives, etc.. At this point it is a stretch, but a chemist studying reactions between molecules could be compared to a linguist working at this level. Based on function in a context a molecule may be called a catalyst, or enzyme, protein, carbohydrate, etc.. So at the end of this set of procedures a linguist has described a phrase such as "Bill kicked the ball" with a full description of all its constituents and, with enough evidence, possibly a rule that expresses a general relation such as "subject-verb-object".

Thus, in Bloomfield's linguistics there are emic units that are built upon data from a lower level of analysis. Starting with the phoneme, we progress to the morpheme. Bloomfield's structuralism requires basic elements to be defined before structure can be analyzed. As one prominent Bloomfieldian linguist pointed out, "if the subject of one's study cannot be so divided up, there can be no technique for describing any possible structure." (Hall, 1978[1972]:86). Bloomfield defines many more "-emic" terms besides phoneme and morpheme, but in practice these two were the core of structural elements in his methodology. Curiously though, Bloomfield defined the term "sememe" as "the meaning of a morpheme" followed by the statement that "the linguist assumes that every sememe is a constant and definite unit of meaning ... but he cannot go beyond this." (1933:162) One wonders why Bloomfield even went that far since it seems to define his conception of meaning as being as rigidly objective as his study of form.

Regardless, it was assumed then that the complexities of syntax would keep linguists busy for a very long time. Semantics, or meaning was not the next step; Bloomfield never proposed that linguistics could one day take data from a robust syntactical analysis and use it on an as-yet-undefined semantic level of analysis (though some linguists did propose a structural semantics after syntax got rolling along, cf. [Katz & Fodor 1963]). While much of the intricacy that made linguistics a challenging and attractive field for scholars in Bloomfield's time has not been mentioned here, it is essentially accurate to characterize linguistic methodology in Language as a series of steps from raw phonetic data, to phonemic, to morphological, to syntactic.

This seems somewhat odd. It is hard to imagine even an unusually descriptive chemist who would take no active interest whatsoever in extending his research to the possible application of his descriptive theories - leaving his data to nutritionists and chemical engineers to interpret as they will. But recall that Bloomfield explicitly left the aspects (A) and (C) in his story of Jack and Jill - the reasons, reactions, and interpretations of the language utterance - outside of the domain of linguistics.

In the division of scientific labor, the linguist deals only with the speech-signal (r........s); he is not competent to deal with problems of physiology or psychology. The findings of the linguist, who studies the speech-signal, will be all the more valuable for the psychologist if they are not distorted by any prepossessions about psychology. (Bloomfield 1933:32)

Any reader with a high-school knowledge of science will have noted that the analogy made here between descriptive linguistics in Language and descriptive chemistry has been mildly stretched in places. However the point should be self-evident. Bloomfield, in his desire to pattern linguistics on the natural sciences, in his mechanistic approach which disallowed subjective input or assumptions to taint the "objectivity" of the observer, created a linguistic science in which sound waves were treated as physical objects. Bloomfield studied language as an object. Reified to the point that language as a process was not a concern of linguistics.

To determine why meaning was never examined in Bloomfield's descriptive linguistics we need not even consider his rejection of consciousness, that is his behaviorism, or his ambiguous statements concerning what meaning might mean. His methodology ensured that semantics could not fit into his descriptive program. Thus, Bloomfield's interpretation of what the scientific method required determined the questions he was willing to ask. This begs the question: is it appropriate to treat a uniquely human, uniquely social, phenomenon such as language as though it was scientifically and methodologically no different than an arrangement of atoms?

Bloomfield's statement at the top of this section, that "the study of language can be conducted without special assumptions so long as we pay no attention to the meaning of what is spoken," must not be overstated to mean that he believed meaning had no relevance in the study of language. But he did have a strong skepticism about how language might be included in a scientific enterprise.

The statement of meanings is therefore the weak point in language-study, and will remain so until human knowledge advances very far beyond its present state. (1933:140)

This was part of both his methodological restraints, and his "psychological prepossessions" which he claimed not to have, but are of course impossible to not have. So what was Bloomfield's "semantics"? Or, considering his contention that semantics might one day develop given an advanced state of human knowledge, what would Bloomfield's semantics become? It might be useful to examine a lengthy quotation from his chapter on the subject of meaning (1933:139-157).

The study of speech sounds without regard to meaning is an abstraction: in actual use, speech-sounds are uttered as signals. We have defined the meaning of a linguistic form as the situation in which the speaker utters it and the response which it calls forth in the hearer. The speaker's situation and the hearer's response are closely co-ordinated, thanks to the circumstance that every one of us learns to act indifferently as a speaker or as a hearer. In the causal sequence

speaker's situation --> speech --> hearer's response,

the speaker's situation, as the earlier term, will usually present a simpler aspect than the hearer's response; therefore we usually discuss and define meanings in terms of a speaker's stimulus.

The situations which prompt people to utter speech, include every object and happening of their universe. In order to give a scientifically accurate definition of meaning for every form of a language, we should have to have a scientifically accurate knowledge of everything in the speakers' world. The actual extent of human knowledge is very small, compared to this. We can define the meaning of a speech-form accurately when this meaning has to do with some matter of which we possess scientific knowledge. We can define the names of minerals, for example, in terms of chemistry and mineralogy, as when we say that the ordinary meaning of the English word salt is 'sodium chloride (NaCl)', and we can define the names of plants or animals by means of the technical terms of botany or zoology, but we have no precise way of defining love or hate, which concern situations that have not been accurately classified -- and these latter are in the great majority. (1933:139)

The above passage is somewhat characteristic of Bloomfield's comments on meaning. It is ambiguous. It makes it very difficult to determine what "kind" of semantics Bloomfield might develop. In his Ph.D. dissertation, Don Mottershead (1989) maintained that there are three possible interpretations from the above passage: a behavioral semantics, a translational semantics, or a referential semantics.

Bloomfield may have proposed a behavioral semantics where the meaning of the form can be defined by the human behaviors that are associated with its use - the stimulus and response. This would be supported by Bloomfield's statements concerning Jack and Jill, where he said the meaning could be defined as events A and C, the observable actions of the speech participants.

Another possibility is a translational semantics in which a form is defined "by translating it into some canonical language, specifically the language of advanced science." (Mottershead 1989:86). An example is when Bloomfield stated that salt may be defined as "NaCl". Bloomfield's text supports this interpretation later with the statement, "...we define the meaning of a linguistic form, whenever we can, in the terms of some other science." (1933:140) A translational semantics also seems to be indicated by his insistence that scientific knowledge must be advanced before we can give a rigorous definition of the speaker's situation or stimulus.

The third choice, that Bloomfield proposed a referential semantics, is also quite possible. In this scheme, the meaning of a form would be defined by the object to which it refers. Bloomfield stated that one may learn the meaning of apple by demonstration (1933:140), and he also stresses the connection between form and referent in his semantics theory. However, this interpretation is muddled by ambiguity. For instance the following quotation could give support to a referential interpretation or a translational one.

Nevertheless, it is clear that we must discriminate between non-distinctive features of the situation, such as the size, shape, color, and so on of any one particular apple, and the distinctive or linguistic meaning (the semantic features) which are common to all situations that call forth the utterance of the linguistic form, such as the features which are common to all the objects of which English-speaking people use the word apple. (1933:141)

Mottershead claimed that these three interpretations of Bloomfield's semantics are not necessarily compatible with each other or with the rest of Bloomfield's work. So we find different scholars making vastly different interpretations of Bloomfield. Mottershead noted that if we can define Bloomfield's interpretation of meaning as behavioral, "then it is difficult to see why it is necessary to have tremendous scientific advances in all sciences prior to doing semantics" (1989:86). The same problem applies to the referential approach. Why is a great scientific advancement needed to define forms by objects, when "the sentence '"Salt" means that white stuff' states exactly the same relation as '"Salt" means NaCl." (1989:86-7)? Aside from these apparent conflicts I do not believe it possible to compartmentalize these three interpretations as essentially separate choices - since none is satisfactory by itself or satisfactorily excluded. Instead, I would like to consider them as concurrent aspects of Bloomfield's semantics and analyze them in turn.

Mottershead concluded his analysis by stating that "the most charitable view of Bloomfield's semantics is that it is intended as a translational semantics supported by behavioral evidence." (1989:87). In analyzing translational semantics he pointed to the biggest proponent of translational theory W. V. O. Quine (1960) and Quine's detractor Donald Davidson (1984). "Quine does not demand that the target language of the translation be the language of advanced science, although he does require that it be properly 'regimented', that is, cleared of many of the vagaries and ambiguities associated with unregimented natural languages." (Mottershead 1989:87-8) So basically in a translational approach to the study of meaning one would have a subject language translated into a non-natural object language through a theory written in a 'metalanguage'.

Of this approach Davidson (1984) had the objection that, "a method of translation deals with the wrong topic, a relation between two languages, where what is wanted is an interpretation of one." (1984:129) Davidson made the case that "if a subject language happens to be identical with the language of the theory, then someone who understands the theory can no doubt use the translation manual to interpret alien utterances; but this is because he brings to bear two things he knows and that the theory does not state: the fact that the subject language is his own, and his knowledge of how to interpret utterances of his own language." (1984:129).

Mottershead noted that Davidson was arguing that translational semantics provides only a postponement of the problem of interpretation or understanding (1989:89). This is a valid comment and it does serve to discredit translational semantics as an adequate theory of meaning. However, I believe that Charles Taylor (1985[1980]:248-92) made even stronger points against Quine, translational semantics, and, if we accept Mottershead's interpretation of Language, Bloomfield's conception of meaning. Taylor attacked the conception of the possibility of an objective observer as well as representational assumptions of how meanings are related (1985[1980]:248-92). I shall return to these issues in Chapter Four.

It would be fine to assume that Bloomfield makes a pitch for translational semantics, discredit it, and dismiss his entire semantic theory. However, because evidence exists for other aspects, I am compelled to look at them as well. Robert A. Hall, Jr. (1972 & 1987a) has argued that the reason Bloomfield could not achieve a satisfactory semantic theory was because he relied too strongly on a referential signal-signifier relation. Hall's opinion is worth some weight in this matter; he was one of the preeminent Bloomfieldian linguists. He notes that the shift in Bloomfield's psychology had a huge effect on his semantics.

In his 1914 book, Bloomfield had recognized the basic associative nature of meaning - the tie between a linguistic form and what the speaker perceives it as referring to. In 1933, he treated meaning as something neither wholly intrinsic nor wholly extrinsic to linguistic form, without a really clear definition of what meaning is or how it is related to linguistic structure. (1987a:156)

When he relied upon Wundt's psychology Bloomfield could accept a 'mental' aspect in his conception of meaning. In fact his Study of Language (1914) even features a chapter entitled "The mental basis of language" (1914:56-71). What Hall points out is that because Bloomfield, in the years between his two books, dropped the terms 'mental' and 'mind' from his vocabulary, he could no longer achieve a stable conception of meaning.

The basic defect of, not only Bloomfield's, but virtually all other treatments of meaning, in linguistics, psychology, and philosophy, is the absence of specification as to where meaning is located. As I have tried to point out on several occasions, we can deal with the problem of meaning only if we abandon the old mediaeval and Saussurean dichotomy of signans and signatum (signifier and signified, etc.), and replace it by the trichotomy suggested by Ogden and Richards (1923) of linguistic form - sense - referent. (1987a:157)

Since Bloomfield eliminated the individual as an intermediary for meaning he had no recourse but to ignore the sense of Ogden and Richard's equation and rely on a basic linguistic form - referent model. Bloomfield's mechanistic philosophy was so strong that by Hall's personal account, "he acquired a psychological block" (1987a:158) against even a scientific or neurological conception of the terms 'mind' or 'mental'. So whether Bloomfield indicated the referent to be an object or a situation, it was not a 'representation' or 'sense' in the speaker's or hearer's mind.

Hall argued that Bloomfield made a fatal mistake in his semantic theory by eliminating the mental aspect in his adoption of Weiss's mechanistic psychology, which made it impossible for him to locate or even consider meaning. Hall's argument brings an important related aspect to our attention. He pointed out that the associative tie in the three-part system can exist "only inside each individual, as part of his linguistic competence. ... Only inside his brain and nervous system [which Hall maintained as the only realistic operational location of mind -krh]" (1972:87). It is not only Bloomfield's disavowal of mentalism that is the problem with this aspect of his semantics, but his insistence on ignoring the individual and objectifying language.

In his linguistics he ignored individual differences in language production, idealizing it, and, as was shown above, reifying language to the status of an object. Of course this is true of his methodology and scientific approach to language only; he did recognize the existence of individual differences, but for linguistics he did not consider them important. This decision was probably justified in Bloomfield's case, given that his primary focus was the creation of a rigorous set of procedures to define linguistic form. However the implications of this on his semantics were enormous. By making an objectivist criterion basic to his approach he had to take an objective approach to his theory of meaning. This is a major reason for Bloomfield's constant insistence that a study of meaning is impossible until scientific knowledge advances. A strongly objectivist descriptive approach naturally presumes that meaning is "the weak point" in language study.

The third aspect in terms of which we may consider Bloomfield's theory of meaning is that of behavioral semantics. This would be one in which the meaning of a form is defined by the behavior associated with its use, including perception. This is probably the closest aspect to Weiss's philosophy of language. Recall Bloomfield's story of Jack and Jill. Bloomfield labeled the stimulus of Jill being hungry and seeing an apple as (A), her saying "apple" as (B) and Jack's reaction as (C). He then makes a blunt statement about meaning.

"...the meaning consists of the important things with which the speech-utterance (B) is connected, namely the practical events (A and C)" (1933:27)

As we saw above, Mottershead saw behavioral evidence as only supporting a translational approach. The dismissal of behavioral semantics as the central conception of meaning in Language rests upon an assumption that Bloomfield employed a simplistic behaviorism, which would require advancement in only behavioral psychology's notions of stimulus and response. There is evidence that this is the case in Bloomfield's story with its simple "s......r" diagram. However considering the influence of the behaviorist Weiss on Bloomfield makes this matter one worth considering. Weiss's definition of all human activities on strictly mechanistic terms made his behavioral description, and hence Bloomfield's, contingent upon the facts of the natural sciences. As Esper pointed out:

The aim was the same as that of certain 'logical positivists,' e.g. Carnap (1959) who declared that 'every sentence of psychology must be formulated in physical language' ... The point of view was that of the contributors to the International Encyclopedia of Unified Science; these included Bloomfield whose contribution [was], "Linguistic Aspects of Science," ... (1968:178)

Esper quoted a passage from this essay (from the International Encyclopedia of Unified Science) in which Bloomfield defined behaviorism as the notion that "science shall deal only with events that are accessible to any and all observers"; mechanism is defined as "only events which can be placed in the co-ordinates of time and space"; physicalism is defined as the restriction of science to a set of terms "concerning physical happenings" (originally from Bloomfield [1939:13], reprinted in [Esper 1968;179]).

Bloomfield's insistence on explaining acts as behaviors in strictly physical terms means that his insistence of advancement of scientific knowledge would have been justified for him even if he did not make the translational aspect of his semantics central. There is evidence for the argument that Bloomfield's conception of meaning was primarily a behavioristic one, defining meaning as the complex situation which acts as a stimulus for a speaker as well as the reaction of a hearer.

However, in the behavioristic aspect Bloomfield runs into the same problems as in translational and referential semantics. Because he wants to include the conception of a sememe with an assumption that it has a "constant and definite meaning", he would have to eliminate situational variation from his semantic theory. However, the whole idea that any morpheme can have a constant meaning is impossible if context is brought into the picture by the inclusion of situational factors. Bloomfield acknowledges this in his discussion of "displaced speech" (1933:141-2) in which he said that lying, irony, jesting, etc. are functions based on the "primary or dictionary meaning" of a term.

How is a linguist supposed to make objective observations and strictly empirical conclusions about the meaning of an utterance if he must reference it to assumptions about a "primary or dictionary" meaning of a term? The matter seems far from clear. Furthermore, because Bloomfield's strict mechanism relied only upon physical 'laws' it could not include meaning as an aspect of human understanding. Thus, defining meaning descriptively, in terms of scientifically defined operations, seems to make behavioristic semantics a type of translational semantics.

The problems with the translational, referential, and behavioristic aspects of Bloomfield's semantics show clearly why he considered meaning to be the weak link in language study. He maintained that linguists should not distort their science with prepossessions about psychology, but what then was his almost vehement dissociation with mental terminology if not a psychological presupposition? The truth is that linguistics in Language was distorted by Bloomfield's own "psychological prepossessions", as any theory of any human activity necessarily is.

The assumptions that language can be studied objectively, that there can be no mental link between form and referent, and that the meaning of a form can and should be defined in behavioral terms, are all strictly speaking tied to Bloomfield's psychological doctrine whether expressed as behaviorism, mechanism, or objectivism. It is naive to assume that because one has become ultra-empirical then one has lost psychological prepossessions. Is the assumption of an empirical viewpoint not a prepossession itself? This is to say nothing about the idea that there is only one true type of empiricism?

Bloomfield's theory of meaning, as well as his entire linguistic program was affected by his assumptions: his insistence on studying language, in the manner of the natural sciences, beginning with formal features such as the phoneme and realizing a language as a reified object; his disavowal of linguistic performance and individual variation as a significant aspect in language science; his insistence that linguists be completely objective in their approach, withholding their intuitions in studying language "without prepossessions"; and his assumption that a successful analysis is even possible through purely objective observation. I will return to these matters in Chapter Four.

Consider the following quotation from Bloomfield.

To put it briefly, in human speech, different sounds have different meanings. To study this co-ordination of certain sounds with certain meanings is to study language. (1933)

Does Bloomfield ever actually define a study of the co-ordination of sounds with meanings in Language? Recall that the quotation at the beginning of this section warned that the only way to objectively study language is to "pay no attention to meaning". The elimination of meaning seems more reflective of Bloomfield's approach to the study of language than of any coordination. But what of the above quotation, which seems so un-Bloomfieldian, in light of the evidence we have just examined? The alert reader will have observed that it is missing a page reference - this is because it was quoted indirectly from another book. Ironically, that other book was a bestseller, making this quite likely the most widely read passage ever written by Bloomfield. The book is Language in Thought and Action by S. I. Hayakawa (Hayakawa 1978:32).

Language in Thought in Action was first published 1939 and was intended as a popularization of the language theory contained in Alfred Korzybski Science and Sanity, to which we now turn our attention.

Previous | Contents | Next