History of Phonological Theory 1950 - 1990
John Goldsmith
Fall 2004: MWF 10:30-11:20


General resources:

1. First person singular: Reminiscences by linguists working during the 1950s
2. Wallace Chafe's thoughts: Searching for meaning in language: a memoir.
3. Goldsmith, ed: Phonological theory, the essential readings. Blackwell's, 1999. I'd like to strongly encourage students to read this book -- the editor has succeeded in putting together an excellent collection (IMHO!).
4. Hymes and Fought 1981 American Structuralism, reprinted from (1975) Current Trends in Linguistics 13: Historiography of Linguistics. The Hague: Mouton.This is an outstanding and thoughtful work that very much deserves to be read.

C1 Overview
General remarks: all significant changes are viewed by their adherents as liberating for the average working linguist (AWL). They allow AWLs to do things they couldn't do before -- sometimes by introducing new tools, sometimes by breaking down rules imposed by dominant perspectives, and usually by doing both.

This course is not designed as a greatest hits of the 50s, 60s, 70s, and 80s; there are some of those hits here, but it's designed to allow you to read papers both from that period and from contemporary phonology better, and better equipped to understand the problems that are being tackled.

C2 Goldsmith and Laks overview, with examples.

Major sea change with the rise of generative grammar:
a. The rise of algorithmic explanation. Its origins in the notion of "mechanical" explanation, and the dislike for "mentalism". A reorientation with the rise of generativism that sees mechanical and mental not as in conflict but as linked. From the algorithm to generative grammar and back again. The variety of modes of explanation that have been important in linguistics: historical, psychological, sociological, algorithmic.
b. A change in relationship to data and to prior scholarship.
c. A change in the view of the relationship to the individual and to the mind.

SPE model: Phonological representation as a linear sequence of segments, defined as a set of feature-values; underlying representation related to surface representation with a derivation, and a set of linear

1. Charles Hockett

C3. Principles of phonologic analysis: excerpt from A Manual of Phonology (1955), pp. 143-180. Seminar discussion paper by Ilya Yakubovich.
C4. A note on structure. IJAL 14.269-271 1948.

Some remarks of Hockett's on studying the history of linguistics (from First Person Singular, linked supra):

I am concerned with archiving because I am interested in two kinds of history of linguistics: more important is the intellectual history, which has to do with who proposed what notion when, who took it up and developed it, and so on; but also relevant is what we might call the "personality" history, which is what we have been speaking of mainly at these sessions. Archiving is the storing up and preserving of the materials from which both sorts of history (and any other kind, for that matter) can be written. I think that the rescue of records is especially important right now because of what has happened to our profession in the last two decades. I shall say nothing here, either favorable or unfavorable, about the quality of research and theory in those two decades. But I do view as genuinely tragic the success of the "eclipsing stance" of the transformational-generative school. We have currently in our ranks a large number of young people, many of them very bright, from beginning students up to and including a few full professors, who know nothing of what happened in linguistics before 1957, and who actually believe (some of them) that nothing did happen. If we lose the records, we lose all chance of recovery from that collective amnesia.

I have rather more sympathy for our disinherited youngsters than I used to have, because of a recent experience of my own. Remember that I cut my professional eye teeth on Bloomfield's book back in 1933. Bloomfield himself assumed no "eclipsing stance": the very opposite, for his respect for his predecessors was profound and he tried to inculcate the same attitude in his students. But I found Bloomfield's synthesis so satisfying (except in some minor technical details) that for a long time I simply couldn't bring myself to read much of the work of those predecessors. That was the price I paid for my largely superb induction into our discipline. Then, just a few months ago, I finally had reason to undertake a serious study of William Dwight Whitney's general writings. I knew that Bloomfield had overtly acknowledged his debt to Whitney; nevertheless, I was overwhelmed to discover the extent of that debt (and thus of our own), and amazed at the variety of topics on which Whitney's remarks, allowing for a difference of terminology and style, are as valid and profound now as a'century ago.

Should my mentors, back in the 1930s, have insisted that I work my way through Whitney? Perhaps so, and perhaps that would have made me a better scholar. On the other hand, possibly I would not yet have been mature enough to tune my twentieth-century ears to his nineteenth-century voice. Our receptivities really do change. I'm sure you won't think me facetious if I offer, as another example, the fact that when I tried Milne's Winnie the Pooh first, during my adolescence, it was unspeakably dull, but when later I picked it up to read to my own children it had become poignant magic.

So (to return to our disinherited youngest generation) I guess what we mainly need is patience. Few of us find the work of recent leading figures dazzling; but many of the kids do. It is bound to take time for their eyes to adjust so that they can see other and earlier things. Our archiving is for them and their successors, because continuity and cumulativity really are crucial in science. When they are ready to look, the record must be there for them to see.

Recommended: Charles Hockett, Linguistic elements and their relations. 1961. Language 37(1): 29-53.

2. Zellig Harris 1909-1992


C5. Chapter 7, Phonemes, in Methods in Structural Linguistics (originally titled, Methods in Descriptive Linguistics). Seminar discussion handout by Colin Sprague.
C6. Nevin collection, Review of The Legacy of Zellig Harris, ed. Bruce Nevin, by John Goldsmith

[T]he rise of descriptive, or structural, linguistics often enough was accompanied by aversion to 'philology', as a label for some practices, attitudes, and styles of work, but there is no simple relation of rejection or supercession. Therein lies one of the major positive features of the rise of structural linguistics in the United States. Leading scholars saw their work, not as the denial of the valid content of philology, but as its extension, indeed universalizing, to peoples and languages without philologies of their own. In this extension there arose, and remain, strains and tensions between established European philologies and the newer ones. Work on the new basis challenged received opinions as to languages and cultures, especially 'primitive' languages and cultures; there were no doubt conflicts of role and personality as well. Nevertheless, structural linguistics in the United States developed on the assumption that it helped to complete the impulse that had created the philologies of classical antiquity and Asia. Hymes and Fought 1975, p. 56.


Steven Anderson, American structuralist phonology, Chapter 11 of Phonology in the Twentieth Century (1985).

Bruce Nevin, Harris the revolutionary: phonemic theory.

Webpage about Zellig Harris: http://www.dmi.columbia.edu/zellig/

3. Roman Jakobson; Features; Halle's argument against the phoneme


C7. Roman Jakobson: On the identification of phonemic entities (1949) Travaux du Cercle Linguistique de Copenhague 5:205-213.
C8. E. Colin Cherry, Morris Halle, and Roman Jakobson: Toward the logical description of languages in their phonemic aspect (1953) Language 29:34-46.
Morris Halle: The strategy of phonemics (1954) Word 10:197-209.
C9. Roman Jakobson: Russian conjugation (1948).
C10. Steven R. Anderson "Reflections on 'The phonetic rules of Russian'" Folia Linguistica 34:11-28.

American and European sensibilities:

...Joos(1957): 'the American (Boas) tradition that languages could differ from each other without limit and in unpredictable ways.'... Insofar as the statement is taken to express a radical relativisim, it is mistaken, and it is misleading in its failure to show the other side of the coin, namely, that all languages could be regarded as equal, both in the eyes of lnguistics and of mankind.
Joos presents the statement as contradictory of the universalism of Trubetzkoyan phonology. (It is European ideas, especally those from Prague, Joos says, that were most often called 'strutural' by the Bloomfieldians, who favored 'descriptive' in the early 1940s.) There was indeed insistence, warranted by experience, as to the danger of imposing a priori categores on languages. The tension between an inductive and an apriori emphasis persisted throughout the development of structuralism, and there was some noticeable difference bewteen dominant American and European outlooks in this regard. But Boas (and Sapir, and Kroeber...) did not at all consider that languages could differ from each other without limit. They held definite conceptions of properties, and an analystic frame of reference, valid for all languages....(
Hymes and Fought 1975 57-58)

Hockett 1951 on Trubetzkoy's phonology, in a review of Martinet:

[333:] By and large my impression is extremely favorable. Martinet presents an outline of fairly orthodox Trubetzkoyan phonology, but by no means implies slavish adherance [sic] thereto on his own part. The presentation is as clear and convincing as any I have ever read….p.342: …Martinet's succinct exposition can very well stand as the last and best summary of Trubetzkoyan methods, an excellent point d'appui for moving from the Trubetzkoy phase of the history of phonological theory into the new phase which we are now collectively engaged in developing. Language 27.3, p.333-42 (1951), review by Charles Hockett of Phonology as functional phonetics by André Martinet.

E. M. Uhlenback, on post-Bloomfieldians

...Post-Bloomfieldian linguistics tended to isolate itself from others...This isolationist tendency [of post-Bloomfieldians to ignore European linguists' work] and lack of interest in achievements elsewhere were not easily overcome and remained characteristic also during the later period of generative grammar. (1979)


Recent remarks by Morris Halle on features: a response to Clements' Features and sound inventories (2004)

The European Background of American Linguistics. Papers of the Third Golden Anniversary Symposium of the Linguistics Society of America. Edited by Henry M. Hoenigswald. 1979. Dordrehct: Forist Publications.

4. Derivational approaches to phonology

The roots of the notion of derivation in phonology: Bloomfield's Menomenee morphophonemics and Chomsky.

The notion of the derivation, with a sequential application of ordered rules, is one of the central ideas of the classical generative (SPE) model. We'll look at some of the roots of the idea. There is a link to historical analysis, and to the type of analysis that Charles Hockett termed "item and process (IP)" analysis in the early 1950s. Don't lose sight of the fact that derivation-l accounts of phonological regularities have risen and fallen on the marketplace of phonological ideas; in recent times, OT has argued against it, and somewhat earlier, a lot of energy was put into alternative non-derivation, or non-processual, accounts of phonology. An excellent discussion of the question from the 1940s through the 1960s is in R. Huddleston, The development of a non-process model in American structural linguistics, Lingua 30 1972, 333-384.

C11. Why phonology is different, by Morris Halle and Sylvain Bromberger, in Linguistic Inquiry 20 (1989):51-70.
C12. Pierre Encrevé, The old and the new: Some remarks on phonology and its history Folia Linguistica XXXIV/1-2 2001. Seminar discussion handout by Jen Watson.
C13. Leonard Bloomfield1939 Menomini morphophonemics Travaux du Cercle Linguistique de Prague 8:105-115. Seminar discussion handout by Stefie Kuzmack.


E. F. K. Koerner Remarks on the origins of morphophonemics in American structuralist linguistics Language and Communication 21(1): 1-43. 2003. See also On 'influence'in linguistic historiography: Morphophonemics in American structuralism, Chapter 4 of Essays in the History of Linguistics, by E. F. K. Koerner. Philadelphia: John Benjamins. 2004. See also his Toward a History of American Linguistics. E. F. K. Koerner. New York: Routledge, 2002. Chapter 9.

Consider the following very interesting remark in Language 27.3 (p. 289f) 1951 in a massive review by Zellig Harris of Selected Writings of Edward Sapir in language, culture, and personality:

Process or Distribution. Sapir, however, also used this model of an 'entity as a result of process' within descriptive linguistics proper. Consider, for example, those environmental ranges by virtue of which two sound types never contrast: say the fact that in a certain language no morpheme contains two vowels in succession; and that in any word which contains one morpheme ending in a vowel, followed by a second morpheme beginning with a vowel, a glottal stop is pronounced between these two vowels. When we speak in terms of distribution and classification, we would say that no morpheme contains the VV sequence, and that all morphemes which end in V before consonant or juncture have alternants ending in V? before vowel (before any following morpheme which beings with a vowel). Hence the VV sequence never occurs across morpheme junction, just as it doesn't occur within a morpheme. In contrast with this, Sapir would say that no two vowels could come together (within a morpheme), and that when a particular morpheme conjunction would have the effect of bringing two vowels together a glottal stop comes in as a protective mechanism to keep them apart. This kind of model appears in much of Sapir's grammatical work and in the work of some of his students, as for example in Newman's handsome analysis of Yokuts.[fn].

We can consider this simply as a method of description, an alternative to our present formulations, which we make in terms of the classifying of occurrences. The process model has the advantage of being more dramatic, and often of reflecting the actual historical changes (the inter-morphemic glottal stop may well have been a later development). [fn] It has the greater advantage of opening the way to a more subtle descriptive analysis ---something always dear to Sapir's heart--by giving a special secondary status to some parts of the descriptive structure. For example, we may be missing something when we say innocently that VV does not occur across morpheme boundary (while V?V and VCV do); the V?V which we find there may not be fully equivalent to the VCV which result from morphemes ending in -VC plus morphemes beginning in V- (or from -V plus CV-); for one thing, these VCV alternate with -VC and V- when their morphemes occur separately, whereas the V?V alternative with -V and V-; for another, the frequency of V?V (differently from VCV) may be much greater in those positions where morpheme boundaries can occur than in other positions. [fn: to make this more explicit: Suppose all word-initial morphemes have two or more syllables (vowels). Then the probability of finding ? rather than some other consonant after the first vowel of a word is related simply to the frequency of the medial glottal stop. The probability of finding ? after the second vowel is related to the frequency of the glottal stop (medial and at the end of morphemes) plus the frequency of morphemes which end with a vowel (and of morphemes which begin with a vowel). However, the probability of finding other consonants (not ?) after the second vowel is related merely to the frequency of those consonants medially and at morpheme-end.] On the other hand, the process model ahs the disadvantage of bringing into the descriptive analysis a new dimension--the relations of one distribution to another distribution-- which does not fit well into the algebraic character of the present bald statements of distribution. There is need for further elaboration of descriptive techniques, in order to make room for such refinements among our direct distributional statements.

The Process and its Result. We can also consider the use of the process model as an activity of the linguists who use it; and we can then say that aside from such personality reasons as may have dictated Sapir's use of it, it also occupies a determinate position from the point of view of the history of science. It seems to constitute a stage in the separation of descriptive method both from historical analysis and from the older psychologizing of grammatical forms. The older grammars did not distinguish descriptive from historical statements, so that the history of the glottal stop at word boundary would have been combined with the statement of the absence of vowel sequences there. The older grammars assigned reason for speech forms: people said V?V (with 'intrusive glottal stop') in order to avoid VV which they did not otherwise pronounce. [fn 6: How different Sapir's psychologism is from this will be discussed in Part 3 below. For the moment, it is worth noting that Sapir's grammatical formulations stayed within linguistic categories. In descriptive linguistics he would not say that people inserted a glottal stop so as to avoid the sequence VV, but that the glottal stop constituted, in respect of medial VV, a 'protection' (in cross-boundary position) of that non-occurrence of VV.' The primacy of medial VV over the cross-boundary case is maintained, but in terms of the structure rather htan in terms of people's intervention in their own speech behavior. ] Finally, the older grammars frequently failed to distinguish morphological from phonological considerations, so that the morphophonemic fact about V?V appearing for -V + V- would be given together with the phonemic fact about the absence of VV. The formulations in terms of process give expression to all this while at the same time separating descriptive linguistics from the rest. This is achieved by the dual character of these formulations: the 'process' of protecting the cross-boundary -V + V- yields the 'result' that V?V occurs.

The process section of this formulation takes cognizance of such factors as were brought out by the older linguistics (or by Sapir's interest in descriptive detail); the result section gives the distributional statement as an item in a separate science of distributions. [fn 7: We can say that the use of base forms in morphophonemics--as in Leonard Bloomfield's Menomini morphophonemics, TCLP 8.105-15 (1939) -- is a further step from history of process toward distributional statements..]

Process in Language Structure. The process model led to a characterization of linguistic structures in terms of the types of process involved in them....To speak of internal change and suffixation and the like as occurring under particular environmental conditions is to give detailed distributional statement of morphemes as phonemic groupings. This last can be described as a combining of today's distributional interests with the interest in process of Sapir (and, in morphology, Bloomfield) and various European linguists; it is a direction of development which would be fruitful in the present state of linguistics. It would be fruitful because linguistics has at present one technique for stating the relation of phoneme to morpheme (morphemes are arbitrary combinations of phonemes) and another for stating the general relation of morpheme to utterance (utterances are composed of stated distributions of morphemes). To take greater cognizance of the phonemic composition of morphemes is to come nearer to the direct relation of phoneme to utterances (utterances are composed of stated distributions of phonemes). This goal will presumably never be reached, because there will always be arbitrary elements in the phonemic composition of morphemes. But if we can make general statements about part of this field, as by noting when the morphemes or alternants consists of added new phonemes or of repeated phonemes or of exchanged phonemes, we leave less that is arbitrary and outside our generalized statements.

5. Generative phonology: features, derivations, abstractness, and an evaluation metric (simplicity):
Sound Pattern of English (1968): Noam Chomsky and Morris Halle

C14. SPE: Excerpt in Goldsmith 1997.
C15. Morris Halle: Phonology in generative grammar.1962. Word 18:54-72.
C16. Morris Halle: On the bases of phonology. From The Structure of Language, edited by Jerry A. Fodor and Jerrold J. Katz. 1964.
C17-18. Fred W. Householder: On some recent claims in phonological theory (1965) Journal of Linguistics 1: 13-34, and response by Chomsky and Halle, entitled Some controversial questions in phonological theory (Journal of Linguistics 1.97-138, also 1965) followed by rejoinder by Householder. Seminar discussion handout by Jen Watson.

Note that the striving for formal simplicity, and criticisms of this striving, did not begin in the 1960s: take a look at Einar Haugen and W. F. Twaddell's scathing critique in Language 1942 of Trager and Bloch's "The syllabic phonemes of English" (Language 1941) -- this latter being a paper later considered a classic of structuralist phonology.

"To us, in all sobriety, the processes whereby Trager and Bloch combine sound-types into phonemes appear more akin to artistic composition than to scientific classification; and the pleasure of watching their operations with the material is rather esthetic satisfaction than schoarly conviction. The notion of "pattern" which underlies this whole discussion needs clarification before it can become a useful term. If it is a set of habits that exists in the neural system of speakers, then its effect on the functioning of language must be demonstrated. If it is only a working hypothesis, a "fiction", its value for the study of language must be unequivocally stated....they expound this sovereign "principle of pattern analysis" by quoting, as its classic formulation, an eloquent magnificat to "Pattern," with only the vaguest of indications of the nature or the application of this principle. ...The striving for fewer symbols leads to some curious results...."

McQuown on Zellig Harris:

Economy is an almost fetish-like criterion in deciding many issues. (Review of Methods in structural linguistics by Zellig S. Harris. Language (1952) 28.4:495-503.)

Eugene Nida on simplicity:

I am in complete agreement with the desire to attain simplicity of statement and to reveal the whole structure of a langauge in terms of such statements; but it is possible that too much emphasis on structural simplicity, both of statement and language structure, can give a false impression of simplicity and can actually misrepresent some of the pertinent facts. The objections which this paper raises are not directed against any lack of logic in the presentations cited. Bloch and Hockett have been thorougly consistent in following to the end the implications of their basic premises. Any disagreement with their treatments is not a criticism of their handling of the data as such, but is directed rather at the fundamental principles upon which all of us as descriptive linguists have been working. (Eugene Nida, The identification of morphemes, Language 24.4:414-441, p. 414 (1948).)


Steven Anderson, Generative phonology and its origins, Chapter 12 of Phonology in the Twentieth Century (1985).
On the revolution in linguistics: some critical comments by Konrad Koerner: The anatomy of a revolution in the social sciences: Chomsky in 1962.
Paul Newman, The reality of morphophonemes. 1968.Language 44.3.507-15.

6. Abstractness, too much of a good thing? & natural phonology & natural generative phonology:

Theo Vennemann, Joan (Hooper) Bybee, David Stampe, Paul Kiparsky

"Some babbling babies probably produce all the sounds ever employed in any language; despite this, every language has a limited number of phonemes. In a language that has only about a dozen phonemes, like Fox, there will be more free and positional subphonemic variants than in a lanaguage with four dozen phonemes. When a child stops making random sounds (babbling), and limits himself to the phonemes used by older people of his speech community, he may be said to have learned to speak his native language. There is no comparable measuring-point in a child's progress in learning the morphemes of his native language." C. F. Voegelin. Distinctive features and meaning equivalence. Language 24.1.132-135. (1948).

C19-20. Critique of classical generative phonology (SPE): Paul Kiparsky: How abstract is phonology? Manuscript 1968.
C21-22 . Natural phonology: Patricia Jane Donegan and David Stampe: The study of natural phonology. (1979) In D. Dinnsen (Ed.), Current Approaches to Phonological Theory (pp. 126-173). Bloomington: Indiana University Press.

La phonologie générative naturelle et la phonologie naturelle, by Bernard Laks. From History of the Language Sciences/Geschichte der Sprachwissenschaften/Histoire des sciences du langage, ed. by Sylvain Auroux, E.F.K. Koerner, Hans-Josef Niederehe and Kees Versteegh. Berlin/New York: Walter de Gruyter.

Explanation in phonology: a brief history of ideas. Juliette Blevins. Ms.

8. Autosegmental phonology, metrical phonology

1. Roots in Zellig Harris' long components (Prelude, in Goldsmith 1976).
3. Roots in Firthian prosodic analysis

C23. John Goldsmith The aims of autosegmental phonology. 1979. In Current Approaches to Phonological Theory, ed. D. Dinnsen, pp. 202-222. Bloomington: Indiana University Press.
C24. J. R. Firth Sounds and prosodies. Transactions of the Philological Society 1948, pp. 127-152.

Response by A. A. Hill: Suprasegmentals, prosodies, prosodemes. 1961. Language 27.4.
John Goldsmith: A note on the genealogy of research traditions in modern phonology. 1992. Journal of Linguistics 28:149-163. See also Disentangling autosegments: a response. Journal of Linguistics 30: 499-507.


Steven Anderson, Generative phonology after The Sound Pattern of English, Chapter 13 of Phonology in the Twentieth Century (1985).

Of interest:

Goldsmith, John. 1973. Tonemic structure (ms.).

Goldsmith, John. 1974. Autosegmental phonology (ms.)

9. Lexical phonology

Morphophonemics redux: abstract phonemes and concrete phonemes. Bringing back into phonological theory a difference between rules that change one phoneme to another (a phoneme??), and those that (without making reference to morpheme identity) account for the range of the variation of realization of a phoneme.

David Stampe, Dorothy Siegel, Mark Aronoff, Steven Strauss, Paul Kiparsky, David Pesetsky

C25. Origins and basics of lexical phonology: Douglas Pulleyblank Tone in lexical phonology, chapter 1.

10. Constraints, phonotactics, degrees of well-formedness, conspiracies

C26. Alan Sommerstein: On phonotactically motivated rules. 1974.
C27. Kisseberth, Singh.
C28. Goldsmith, Harmonic phonology.
C29. Optimality theory

Hymes and Fought 1975: Some recent work on 'conspiracies' in phonology appears to discover the phenomena dealt with by Sapiran structuralists in such terms as 'canonical formulae', and even to rely for data on Sapirian descriptions. American Structuralism, p. 20.

Some issues not covered:

Search for universals. Some interesting comments in treatment of Joos' famous quote in Roger Bacon and Martin Joos: Generative linguistics' reading of the past. Margaret Thomas. Historiagraphia Linguistica XXIX: 3.339-378. 2002.

Rule ordering. Some comments in The abandonment of extrinsic rule ordering in generative grammar. Víctor M. Longa. Historiagraphia Linguistica XXVIII: 1/2.187-198. 2001.

Notes on ideology.

Boas and the development of phonology: comments based on Irogquoian. Paul M. Postal 1964. IJAL 30(3):269-280.


Student papers:

Seminar papers: approximately 3 pages long, submitted electronically to me in text, .rtf, or .pdf format, so that I can post them and link them to this course outline. A seminar paper should present the central idea of one of the papers that we look at (or part of the paper, if we have divided the paper up into pieces) ; it should explain what idea that was current at the time is being challenged, why it is being challenged, and what is being proposed as an alternative. Questions to answer: How does the author try to persuade the reader of the validity of the position? Is it presented as a liberating new idea, and if so, how and why?How does it relate to any ideas you know of that are being discussed today?

Final paper: Topic is due in written form by Week 8. Final form is due by the middle of exam week. The final paper should ideally be an expansion of one or more seminar papers. Length: 10-15 pages (double-spaced).


Why do we study the history of this period? One reason is that we learn a great deal from people who are smart and know how to write well, and such people are in short supply: if we limit ourselves to the current journals, the fact is that we will not read a lot of great literature. There will be some, of course, and we also read the literature to learn from reality, so to speak -- to learn more about languages; but we want to learn from our colleagues, and we have to go to where they are, which is often in the past. In effect, I'm offering a guarantee that reading the best literature from the past will mean reading a whole lot of very good literature, written by smart people tackling important questions.

And problems in linguistics (like in any other discipline) are for the most part not questions that simply get answered, once and for all, in such a way that we need only remember the answer and do not need to remember the question. Most of these questions are not like: what is the speed of light, or what is the diameter of the Earth? -- questions whose answers seem (at least at a first approximation) to be simple and straightforward. Most of the question that get asked are tough because there are two, or sometimes more, answers which are appealing and for which arguments or evidence can be adduced. In such cases, it is often the case that the right answer (or what we think of as the right answer) changes over time: that one answer is best in a given context, at a certain moment in the life of phonological theory.

I think that the most exciting reason to study the history of phonology, for the working phonologist (and that includes anyone who wants to write a doctoral dissertation on phonological theory) is to rediscover the passion that lies behind virtually every plank of phonological theory. What gets a phonologist excited about a theoretical proposal is just about always a realization that if we modify the theory in such-and-such a way, then we will be liberated from some shackle of the past, and capable of analyzing some set of data in a new way, a way that seems now far superior to anything we had seen in the past. That's fine; that's as it should be. But that description of the scenario makes it seem like the past is always shackle, the future is liberation, but there's an inherent contradiction in putting it that way: after all, the past was always somebody's future. What appear to us today as shackles from the past which we wish to break down were somebody's liberating moments. Ordered rules and derivations! Binary features! Identical sounds assigned to two distinct phonemes depending on their context! Whatever it is that you want to leap past, there is someone in the past who was there making a strong and revolutionary case for that position. You must go and meet that person (or whole series of people) and listen to what they had to say -- you may discover that you are a counter-revolutionary who wants to go back to an even earlier position, or you may decide you were wrong, or you may simply come to understand your position a whole lot better. But there is no idea that any of us can have today that does not have a background of violent discussion and disagreement that lies behind it, and that is what we can learn about if we read some of the history of the literature of our field.

Another reason to study the history of the field of phonology is to be able to judge and evaluate the ever-present claims that this or that approach to phonology is more scientific, or that finally we have a scientific approach to a subject which we didn't have before. Since the foundation of the Linguistic Society of America in 1924 (and indeed, even before then) linguists have gotten up on their high horse (if you will allow me a small note of sarcasm) to proclaim that for the first time we have a truly scientific approach to the subject at hand. Now, linguists are not professionally trained to know the difference between science and non-science, so one has to take these declarations with a grain of salt, but obviously they are very much heart-felt, and equally obviously, these remarks hit home and have a lot of impact. But we should learn something, should we not, when time and again linguists look up from what they're doing and declare, Finally we've got it right; finally we're a science, which we weren't before -- but never mind, we've got it right now. From here on out, it's smooth going, because we have science on our side. I'll cite a variety of observations of this Messianic sort below. A certain judicious measure would suggest that each of these developments is a continuation in social process which we call linguistics and whose entire evolution in time is, or isn't, governed by the prinicples that make a field a science.

Here's an example from 1949 -- Bernard Bloch's obituary of Leonard Bloomfield:

"There can be no doubt that Bloomfield's greatest contribution to the study of language was to make a science of it. Others before him had worked scientifically in linguistics; but no one had so uncompromisingy rejected all prescientific methods, or had been so consistently careful, in writing about language, to use terms that would imply no tacit reliance on factors beyond the range of observation. To some readers, unaware of the danger that lies in a common-sense view of the world, Bloomfield's avoidance of everyday expressions may have sounded like pedantry, his rigorous definitions likejargon. But to the majority of linguists, the simply clarity of Bloomfield's diction first revealed in full the possibilties of scientific discourse about language. It was Bloomfield who taught us the necessity of speaking about lange in the style that every scitentist uses when he speaks about the object of his research: impersonally, precisely and in terms that ssume no more than actual observation discloses to him."



Hymes and Fought: American Structuralism

p. 20: We do in fact that if the fundamental premise of structuralism is seen as the study of language as an autonomous system, a system central to the understanding of the history and use of language, but to be analyzed independently of history and use first, then the ways in which Chomsky's work continues preceding structuralism and completes it seem more decisive than the ways in which it does not. Or, if we see in the history of structuralism a complex development, with specific complexities at any one time, and the autonomy of linguistic form as an initial impulse and goal gradually arrived at, then again the unfolding and deepening of that impulse and goal would seem to culminate, so far as we can now see, in Chomsky's work. All of the preceding steps in the exploration of autonomous linguistic form are given a place in a single system, and the autonomy of that system from history and use is rationalized, indeed, given the noblest significance possible, by interpreting it in terms of the human mind, of the distinctive powers of human nature.

(1) With many Bloomfieldians, synchronic, static arrangement and classification of the observed came to be identified with a scientific approach, as against representation and imputation in terms of diachonric or dynamic proocess. [Even so, those most active in fostering this outlook often did historical as well as synchronic work.] p. 56

Genres of presentation are of interest here, and the roles of editors and advisors. In the United States certain modes of obtaining and presenting results were developed and communicated without being fully articulated as models, but through being implemented in relations between scholar and colleagues or students. The Boas ‘model’ for American Indian linguistics has attracted attention and analysis (Hockett 1952b:90, Voegelin 1952a: Stocking 1974); the existence of an early Sapir model of grammar has been noted (Hymes 1963:88-90), and a later Sapir model, exemplified in Newman’s Yokuts, has been explicated (Harris 1944). The collecting of texts, the eliciting of supplementary forms, the organization and preparation of these, the care of notebooks, and later of wire and then tape recorders, all these have underlain the existence of a practice of linguistics. The persistence of habits of work is a major factor in the persistence of loyalty to theoretical goals…17-18…The importance of this dimension of the development of lnguistics can hardly be overstated; yet, evidence for its historiography can hardly be found.

To modes of presentation and acquisition of data may be added modes of argument and persuasion. All are linked to general notions, of course, but also to personalities, life styles, climates of opinion, and the like.

Philosophy of science

Note what Halle cites as justification for a theoretical step: From Morris Halle, "The strategy of phonemics" Word 10.197-209 1954:
In sum, both the continuous and the discrete representaiton of speech can--at least in principle--be achieved by a set of physical operations. We can, therefore, assert that both views are meaningful in an empirical, physical sense. Footnote: "A term (predicate) is a legitimate scientific term (has cognitive content, is empirically meaningful) if and only if a sentence applying thet term to a given instance can possibly be confirmed to at least some degree." R. Carnap, "Truth and Confirmation," in H. Feigl and W. Sellars, Readings in Philosophical Analysis (New York, 1979), p.123.

And a young (26 year old) Chomsky wrote:

But it is clear that such an ad-hoc approach to the problem of classification and characterization of elements in particular languages will be of no help to linguists, who are interested in the general grounds by which these elements and relationsa re established in each particular case. If we have an operational account of synonymy or transformation, this formal and rigorous procedure of arbitrary listing is not needed; and if we have no operational account, this procedure, despite its literal formality, its obviously useless.

Logical syntax and semantics, Language 1955. 31(1), p. 38ff.

Sapir on universalism, in his 1922 Takelma:

Not that Takelma is in the least thereby relegated to a peculiar or in any way exceptional position. A more objective, unhampered study of the languages spoken in various parts of the world will undoubtedly reveal a far wider prevalence than has been generally admitted of the inflectional type. The error, however, must not be mde of taking such comparatively trivial charcteristics as sex gender, or the presence of cases, as criteria of inflection. Inflection has reference to method, not to subject-matter. [emphasis added]. p. 54. [Cited in Hymes and Fought, 60.]

Einar Haugen (in First Person Singular):

...but I was repelled by the attempts of the Blochs and Tragers and Harrises to fit language into procrustean theories which accounted for only a tiny fraction of the linguistic reality I knew. I have in one paper (1972) distinguished between 'procrustean' and 'heraclean' linguists, the latter being the ones who dig into the muck of the Augean stables and try to clear it up without having a preconceived plan of what they are going to find. A student of mine got even by dedicating her book to me as 'the theory-minded Heraclean' (Henriksen 1976). In my presidential address of 1951 I called for a common meta-language for linguistics, one which would unite the metadialects then competing for attention. Six years later a new one emerged, and I welcomed it until I discovered that Halle and Chomsky were merely replacing an old dogmatism with a new one. Happily his followers are falling into the familiar pattern of schismatic rejection. Hegel was right in seeing history as an endless series of rejections of the past, or in Freudian terms, the killing of the fathers and the rediscovery of the grandfathers. Every time I am annoyed by the arrogance of the current generation I think back and recall that some of us were just as arrogant toward our predecessors.

On The University of Chicago

...[A]n even more fundamental revision of linguistic outlook came my way in the 1930s while I was working on the University of Chicago campus. In 1932 I went there as an editorial worker on the Dictionary of American English, under the dirction of Sir William Craigie, who had transferred there from Oxforrd University. His austere scholarship has remained an ideal throughout my life.

My lexicogrpahical work was fascinating in dealing with American cultural history, but it involved very little linguistic theory. Such theory was provided from other active centers on the campus. Intellectual interests were very much polarized by the new young president, Robert Maynard Hutchins, and by his henchman Mortimer Adler. They espoused the "Great Books" movement, which seemed to many of us to be backward-looking and in opposition to experimental science.

The "Unified Science" movement had gained a strong foothold there, much strengthened by the refugees from the Vienna Circle, who had fled from Hitler's advance -- especially Otto Neurath and Rudolph Carnap. The University of Chicago Press accepted the project of the Encyclopedia of Unified Science, with contributors working on the frontier of a new scientific outlook, which they called "physicalism."

I believe that this represented a genuine scientific revolution -- what is now called, in Thomas Kuhn's term, a new paradigm. Its effect was to deny the dualism that had plagued Wsetern thought for millenia. In place of a physical realm plus a mental relam, a single, unified realm was postulated, and the so-called mental realm was explained as the working of the abstracting process, developing out of and on the physicalist base.

Already at work on the Chicago campus were the followers of George Herbert Mead, who had made the symbolizing process the center of his work. Also we kept hearing about the works of Charles S. Pierce, which were being edited on the campus, although published at Harvard. A chief exponent of the new paradigm was Charles Morris, who was developing the field of semiotics. His monograph of 1938, Foundations of the Theory of Signs, established a firm base, and this was built upon in 1946 by his replete, masterly volume, Language, Signs and Behavior. This movement is sometimes called "behaviorism," but that much-abused word does not do justice to the fullness and richness of the unified science outlook.

This background was thoroughly congenial for a newly developing linguistics. In the Antrhopology department, following the lead of Columbia's Franz Boas, Manuel Andrade was develoing the new paradigm; but most importantly, Leonard Bloomfield had come in 1927 from Ohio State University....

Seminar meetings, open to the University community, were held on Friday afternoons, at which advanced thinkers presented their new formulations, and the atmosphere of intellectual ferment was strong. At Hutchinson Commons, where we got a good dinner for forty cents, there would be heated discussionsa t the tables fo graduate students.

I regarded Bloomfield as so important, not because he said the last word in linguistic method, for he didn't. There was plenty left to do for Kenneth Pike, for Henry Lee Smith, for Bernard Bloch, for Charles Hockett, for Eugene Nida, for Robert A. Hall, for Allan Gleason, for Dwight Bolinger, for Uriel Weinreich, and for the many, many others who have produced fresh and stimulating formulations. (The individuals I have mentioned are those that have especially influenced me.) But Bloomfield brought into linguistics the physicalist paradigm that makes scientific rigor possible. Many linguists still work under the tired old dualism, and a genuine rigor cannot be expected from them....

On 165 Broadway:

When the war reached American in 1942, I was drafted into the Army and was so fortuante as to be assigned to a dictionary-making outfit, with offices in a building overlooking the Battery on the lower tip of Manhattan. We compiled materials fora dictionary fo military terms, under the direction of two competent civilians, Leonard Bacon and Grover J. Cronin, with a Colonel in administrative charge. This Colonel was something of a publicity hound, and got our project reported on the front page of the Sunday New York Times. This incensed his superiors in Washington so much that they had our dictionary declared a "military secret." Thus it became a crime to mention what had appeared in the preceding week on the front pgae of the New York Times...

Then in 1943 I was especially lucky to be assigned to the Language Section of the War Department at 165 Broadway, where various linguistic projects were being tried out under the direction of Major Henry Lee Smith. I lived on a "quarters and rations" allowance in Greenwich Village, and felt that it was almost indecent to be in such an interesting and stimulating situation when a war was going on. Henry Lee Smith and George Trager were getting together their system that resulted in the book of 1951, An Outline of English Structure....

On Roman Jakobson:

We had many shorter term workers, who stayed for varying periods, such as Stanley Newman, Charles Hockett, Fred Householder, Paul Garvin, Herbert Penzl, William Gedney, Raven McDavid, and many others. We felt that we were carrying on an American-based linguistics and were not cordial to the intrusion of certain refugee scholars. This was resented by some of them, who felt that they were superior to American scholarship. Especially difficult to deal with was Roman Jakobson, who seemed to us at that time to be overbearing and self-aggrandizing. Bloomfield, with his characteristic humility, in an interview listened to him patiently, but Jakobson later boasted, "I set him right on many matters." We did recognize Jakobson's high worth, and some of the linguists with pull managed with great effort to get him a position at Columbia. Soon afterwards he was offered a better position at Harvard and left New York without thanks for th efforts that had been made in his behalf.

It has surprised me to find that in a paper by Thomas Sebeok, published in 1977, our office at 165 Broadway is described as being a "cabal" and a "powerful clique" consisting of "misguided chauvinists". Sebeok's criticism is as follows:

Jakobson's teachings...were roundly condemned by an influential cabal of autochthonous and lately-naturalized linguists -- mostly a generation or two older than mine -- clustering around "165 Broadway."...In truth, these men were mostly misguided chauvinists, afflicted with a hubris doubtless induced by pressures and fears of an uncertain military conflict in the backdrop. Regrettably, the behavior of this small but powerful clique -- which caused Jakobson and his friends untold anguish, to say nothing of economic loss -- left a sinister stain on the otherwise magnificent tapistry of achievements of American linguistics in the 1940's. Fortunately, this dark episode was transpierced by brilliant shafts of light emanating from giants like Boas and Bloomfield; their instant appreciation for Jakobson's decisive presence must be allowed to compensate for al the rest, which had better stay buried with other, similarly motivated, wartime debris. ["Roman Jakobson's Teaching in America," in Roman Jakobson: Echoes of His Scholarship (Lisse: Peter de Ridder Press, 1977), p. 416]

Sebeok paints a picture very different from the one I remember. There was no "chauvinism" whatsoever, but at most a touch of justified "nativism"-- a pride that helps to keep a culture's traditions in a healthy state. At best it was a difficult problem to take care of such an influx of learned refugees, and some feelings were bound to get hurt. Our member Robert A. Hall, Jr., has explained the attitudes of the time in a study published in 1969. He wrote:

However, the strong anti-European feeling of many American linguists in the 1930's and 1940's had its main roots in often-times bitter personal experiences. Not a few young Americans saw, and frequently more than once, positions (for which they had been trained and were eminently qualified) snatched from under their noses and given to European refugees. Such a reaction, though by no means generous, was easily understandable in the days of the depression when any job at all was hard to come by, especially since American scholars, then as now, were not protected by citizenship-requirements of the kind prevailing in virtually all European university-systems. A frequent remark heard from [these American scholars]...was "We'll show those Europeans we have something they never dreamed of!"["Some Recent Developments in American Linguistics,"in Neuphilologische Mitteilungen, LXX (1969), p. 194, footnote 3.]

-- Allen Walker Read [ (1906-2002) ]. A personal journey through linguistics. The Fourteenth LACUS Forum, ed. Sheila Embleton. Lake Bluff, IL 1988. pp. 5-17.

More on Roman Jakobson, from Morris Halle:

Jakobson arrived in the USA in the summer of 1941, in the middle of the second world war. In spite of many attempts, he was unable to obtain a regular university appointment during the war. I know from conversations with Jakobson that he had extensive negotiations with the Universty of Chicago and that Zellig Harris tried at one point to obtain an appointment for him at the University of Pennsylvania. None of these attempts was successful and, as we shall see in a moment, there is some indication that the failures were due to an active effort to block Jakobson from ever obtaining a regular unviersity position in the United States. As a result he spent the war years lecturing at L'Ecole Libre des Hautes Etudes, which was organized at the New School for Social Research in New York. It was only in 1946 that Jakobson succeeded in being appointed to a regular faculty position at Columbia University....

Roman Jakobson himself wrote:

[T]he question of purported hostility between American and European linguists comes to naught. Any actual contact puts an end to the belief that these were two separate and impervious scientific worlds with two different, irreconcilable ideologies. Sometimes we hear allegations that American linguists repudiated their European colleagues, particularly those who sought refuge in this country. I was one of those whom the second world war brought to the Western hemisphere, and I must state that the true scholars, the outstanding American linguists, met me with a fraternal hospitality and with a sincere readiness for scientific cooperation. If there were signs of hostility and repudiation -- and they were indeed evident--they occurred solely on the side of a few inveterate administrators and narrow-minded, ingrained academic bureaucrats and operators, and I am happy to acknowledge the unanimous moral support and defence which came from such genuine men of science as Charles Fries, Zellig Harris, Charles Morris, Kenneth Pike, Meyer Schapiro, Morris Swadesh, Stith Thompson, Harry V. Velten, Charles F. Voegelin, and many others.

One of the first American linguists whom I met on my arrival in this country and who became a true friend of mine was Leonard Bloomfield. Both orally and in writing, he repeatedly expressed his aversion to any intolerance and he struggled against 'the blight of the odium theologicum'and against 'denouncing all persons who disagree' with ones interest or opinion or 'who merely choose to talk about something else' (in 1946). The fact that one, Bloomfield wrote, 'disagrees with others, including me, in methods and theories does not matter; it would be deadly to have one accepted doctrine'(in 1945). I recollect our cordial and vivid debates; Bloomfield wanted me to stay and work with him at Yale, and assured me that he would be hapy to have someone with whom he could have real discussions. The great linguist severely repudiated any selfish and complacent parochialism.


Some relevant remarks in the extremely interesting LSA Presidential address in 1951 by Einar Haugen, Directions in Modern Linguistics (Language 27.3:211-22).

The very growth of an independent group of linguists has promoted a kind of scientific isolation, with even a hint of arrogance, which can only be deplored by those who, like myself, believe that our science should continuet to be international. [fn: Cs. SIL 8.8 (1950), where the label 'unscientific' is applied to 'much of the European structural studies'; ibid. 8.100: 'the usual kind of European philosophizing on the baiss of insufficient evidence'. ] American linguists are finding it increasingly difficult to read European writings in our field; younger linguists are neglecting the older writers, so that we are in some degree losing contact both with the tradition of linguistic science and its present-day representatives in the rest of the world. Rarely does one see a reference in American writings on linguistic theory to the works of de Saussure, Trubetzkoy, or other European writers, although they were the thinkers who gave us the instruments whith which we work. I yield to no one in my admiration for Bloomfield and Sapir; but I regard it as a kind of provincialism to suppose that all sound linguistics began with them.