In Praise of “Exceptionless:” Linguistics among the Human Sciences at Bloomfield and Sapir’s Chicago

Michael Silverstein
University of Chicago

Edward Sapir (1884-1939) arrived at the University of Chicago for Autumn Quarter, 1925, having spent the summer, in transit from Ottawa, in New York City teaching summer school at Columbia. Two years later, in 1927, Leonard Bloomfield (1887-1949), a native Chicagoan who had his Ph.D. in Germanic Philology from Chicago (1909), returned to succeed his Doktorvater, Francis A. Wood (1859-1948), who had just retired, as Professor in Germanic. The two great figures in the history of linguistics in America were thus colleagues at Chicago for four years, through the 1930-31 academic year, after which Sapir removed to New Haven as Sterling Professor and founding Chair of the Yale Department of Anthropology, with a concurrent appointment in Linguistics. (After the death of Bloomfield’s early mentor, Eduard Prokosch, in August, 1938, and of Sapir, in February, 1939, Bloomfield, too, would go to Yale as Sterling Professor in Linguistics and Germanic, in effect replacing both.)

Linguistics at Chicago, originally denoted by the expression “Comparative Philology,” was one of the original subjects, or “departments of knowledge,” filled by President William Rainey Harper (1856-1906). Harper himself had gotten his Ph.D. in Hebrew and Semitic Philology at Yale under the great William Dwight Whitney (1827-1894), and in 1892 another Whitney student more in the master’s image, Carl Darling Buck (1866-1955), started as “Assistant Professor of Sanskrit and Indo-European Comparative Philology” on return from a degree at Leipzig. By the mid-1920s, Buck presided over a department captioned “Comparative Philology, General Linguistics, and Indo-Iranian Philology,” with a set of course offerings both by himself and by various others whose primary appointments were in other departments. Sapir, originally appointed in Spring, 1925, as Associate Professor of Anthropology and of American Indian Languages in the Department of Sociology and Anthropology (as it then was), taught a range of linguistics courses, listed as well among the offerings in Comparative Philology, to the coterie of students who would continue and expand Boasian anthropological linguistics in the 1930s. (They would all be marginalized by the more doctrinaire Bloomfieldians, as it turned out.) By Autumn Quarter 1927, Sapir was a full “Professor of Anthropology and General Linguistics,” and early in Winter Quarter, 1928, Buck formally arranged for Sapir’s titular appointment in the department, as attested by an exchange of administrative letters with higher-ups in the central administration. Bloomfield’s courses in Germanic, too, were listed under the umbrella of Comparative Philology on his 1927 return, and since Francis Wood had been a member of the department, so, too, was his student and successor, Bloomfield (and would be Bloomfield’s successor, George J. Metcalf [1908-1994]). A number of the students of the late 1920s, for example Li Fang-Kuei (1902-1987) and Mary R. Haas (1910-1996) took courses with all three of Buck, Bloomfield, and Sapir, recalling for me in later years their very different pedagogical styles and emphases.

But it is not pedagogical style as such that concerns me so much as intellectual and professional affiliations and outlooks, and the way that these emerged in a face that linguistics as a discipline would for some time show to its congener disciplines at Chicago and in America more generally. Everything we know about Sapir’s and Bloomfield’s biographies shows that their centers of intellectual gravity and their aspirational commitments pulled in very different directions. This was already very clear by the time they had joined Buck’s enterprise at Chicago, each of them in his early 40s and each of them accomplished and authoritative in his respective scholarly work, and each widely connected in their own disciplinary networks of “invisible colleges,” as Merton termed them.

Sapir was the broader scholar with the ability to dazzle an audience of laypersons as well as graduate students with an extempore display of genius at work, integrating details of some matter in a framework of wider significance. By the time he arrived at Chicago, he had published in The American Mercury as well as many other New York highbrow intellectual venues; he had as well compiled an incredible record of descriptive and comparative publication on Native American languages and ethnological phenomena, the vast majority gathered in his own fieldwork (which he continued vigorously at Chicago, for example on Gweabo [Jabo] of Liberia, and Athapascan languages Hupa, Navajo, etc.). Sapir moved easily in the field of Boas’s first generation of anthropology students, focused on problems of typological and historical reconstruction of the culture history of North America as well as on the way that newer currents in psychology and psychiatry – Freud, Jung, Koffka – were disturbing the New York intelligentsia and, disciplinarily, the various emerging social sciences on the North American side of the Atlantic. His essays from this time forward are arabesques of newly traced connections. He wrote poetry and poetry criticism, book reviews, etc.

Bloomfield, too, appeared at Chicago with a distinguished record of fieldwork on Tagalog and on at least three Algonquian languages, Menomini and both Plains and Swampy Cree (funded through a contract from the Victoria Memorial Museum of Canada that was arranged by Sapir). He was an Indo-European comparative philologist by training, to be sure, with a specialty in Germanic and a love of Sanskrit. Both Malayo-Polynesian (Austronesian) and Algonquian were additional specialties he cultivated as he suffered from the ennui of a brilliant analytic intellectual cast as a German, Gothic, and Dutch language instructor during the decade of the ’teens at University of Illinois. Bloomfield saw linguistics as applicable to practical matters, such as language pedagogy and literacy, and his triumph in this was to come later in wartime mobilization of the whole profession of linguistics for the ACLS/Army Language Program.

The two men obviously met intellectually on the field of American Native languages, as Franz Boas (1858-1942) termed them. Both clearly saw that these languages and their descriptive and comparative study contributed something unique to the development of linguistics in America. Indeed, fieldwork with people who were among the last fluent speakers, if not everyday users, of these languages was, for three or four academic generations, a central component of the training and professional activities of linguists in America no matter what their main scholarly specialties. The data from such field elicitation underlay theory and teaching of analytic methods as well. An empirical triumph of those in Boas’s footsteps, this became the prototype for all “inductive” study of languages during World War II and the foundation of the linguistically informed language pedagogies of post-War “direct method” or “spoken language” teaching – even of Latin!

Recall a bit of the organizational context of Sapir’s and Bloomfield’s mutual gaze as Chicago colleagues. From 1910 on, Sapir had been running the Division of Anthropology at the Victoria Memorial Museum, administering collections and supervising exhibits as well as contracting with various others – such as Bloomfield himself in 1925 – as fieldworkers. He had already held various positions in the governing and editorial councils of the American Anthropological Association as these migrated from old line scholars to the up-and-coming Boasian generation, and thus linguistics as such was for Sapir the “subdiscipline” of Boasian anthropology in which he simply had no peer and in which all deferred to his genius. Sapir looked at culture through language-in-use, and thence, at a host of problems in the history and sociology of humankind. He came to Chicago, in fact, for the promise of psychological anthropology as part of a Rockefeller Fund project to integrate the social sciences around problems of “deviance” – we would now term it ethnic and other kinds of diversity – in American society. Chicago’s departments of the social sciences were already at the forefront of such integrative moves as the founding, in late 1923, of the Social Science Research Council as a coordinating consortium of the principal social science associations for the purpose of maintaining a flow of support to cutting-edge research and training at the major centers in the various disciplines. By 1927, Sapir was a member of several of its boards and committees.

Bloomfield, to be sure, belonged to the various philological organizations and attended their meetings, where he gave papers. But decisive in his trajectory was his founding role as the youngest of the triumvirate of George Melville Bolling (1871-1963), Edgar Howard Sturtevant (1875-1952), and himself, the organizing committee that led to the founding, in late December, 1924, of the Linguistic Society of America. Bloomfield wrote the letters to enlist people – among them, by contacting Boas (letter LB to FB 18 I 24), Sapir – to sign on to “The Call” for the organizational meeting, and in a real sense the organization ever after was his baby, the institutionalization of what he again and again, following Bolling, referred to with the welcoming, inclusive expression, “our science,” all the while stimulating and presiding over its exclusionary narrowing in the name of Science. It was, for Bloomfield, clearly a matter of differentiation of “our science” from philology and literature, the domains of the American Philological Association, the American Oriental Society, the Modern Language Association, etc., and from the rest of the social sciences. “In the domain of anthropology –” he would proclaim in plenary address to the MLA & LSA in Cleveland in 1929, “– that is, in the study of man’s super-biological activities – science has been unsuccessful.” (A Leonard Bloomfield Anthology [1970], 227) Except, as it turns out, in the triumphs of linguistics, which, over the course of human history, have been two: Pāņini’s grammar and “the historical linguistics of the nineteenth century” which “owed its origin largely to Europe’s acquaintance with Indian grammar.” (LBA, 219 [Rev. of Liebich]) Thus, as Bloomfield noted, Sapir (who had a year earlier spoken on “The status of linguistics as a science”) “would probably agree with very little of what I am saying tonight” (LBA, 227); indeed, Bloomfield’s stance on what counts as worthwhile “science” was limited to inductive generalizations about linguistic form and its changes over time.

Sapir and Bloomfield both were trained in the methods of Indo-European comparative philology and Neogrammarian historical linguistics. Both were specialists in Germanic, in fact, though it was Bloomfield who was employed professionally as such. Both men were offended that as late as 1924, in Les langues du monde, no less a personage than Antoine Meillet (1866-1936), surely among the greatest Indo-Europeanists – after Ferdinand de Saussure (1857-1913) – could opine that the regularities of Neogrammarian “sound laws” were no doubt suspended on the North American continent, making comparative-historical linguistics there impossible. (Franz Boas, perhaps, had been one unfortunate source of this view, his series of editorials and papers from a diffusionist point of view asserting that Stammbaum concepts of relationship made no sense for indigenous North American languages – a view he continued to hold despite Bloomfield’s and Sapir’s triumphs of method in language family after language family!)

It was on such grounds that Bloomfield had, in fact, formed his exceedingly high opinion of Sapir. His early approaches to the senior scholar of the linguistics of American Native languages, started in early 1924 when, based in Columbus, Ohio at Ohio State University’s Department of German, he wanted Sapir’s support for founding the Linguistic Society of America. As well, he wanted Sapir’s intervention in landing a Canadian government contract to study Cree, one of the “Central Algonquian” languages akin to Fox, on which he had done philological work from others’ published texts, and to Menomini, on which he had done two summers of fieldwork in 1920 and 1921. His earliest note (LB to ES, 29 I [11 II] 24) compliments Sapir on the latter’s American Mercury piece of February, 1924, “The grammarian and his language,” proclaiming it “to be the very best kind of popularizing and at the same time of sound and novel scientific content.”

As the epistolary relationship developed, Bloomfield – who did, in fact, get a research contract for Summer, 1925 – had been drawn into the early production of the Linguistic Society’s journal by its first Editor and Chair of the Committee on Publications, none other than his senior colleague and close professional mentor at Ohio State University, George Melville Bolling. Since Bloomfield was clearly the theoretical head behind the transformation of linguistics into “our science,” Bolling gave him the manuscript of Sapir’s “Sound patterns in language” to read in early 1925, as the second issue of Language was being prepared. In a letter to Sapir dated 9 April, Bloomfield said that the paper “greatly pleased [him]” because it is “really beautiful,” and offered, diffidently, some queries and a substantive improvement to one of Sapir’s points (on the [ŋ] in sing : singer, duly acknowledged in the published text [Sapir 1925:49, n.6]). Later that month, on the 25th or 26th, Bloomfield reports that Bolling “insisted on my reading your S. Paiute grammar, although I don’t see how there could be any question of its value. … [I]t is beautiful, both the language and the work.” By the 6th of May, Bloomfield reports that he “ha[s] told Bolling he should get your texts and lexicon, as well as grammar.”

In this very same letter, Bloomfield reports to Sapir that “Bolling will probably soon send you the paper I have submitted for Language, in this I try to set up equations for the sounds of Men[omini], Fox, Cree, Ojibwa.” This is clearly “On the sound system of Central Algonquian,” printed in the journal’s second volume (pp.130-56) in 1926. Sapir sent comments to Bloomfield, acknowledged and responded to in a letter of May 27th 1925. Again on the 3rd of June, 1925, Bloomfield continues a responsive discussion of his manuscript, among other matters, and really begins to open up in the tone of what we recognize as an enthusiastic and grateful structurally junior person. Bloomfield’s discussion turns to general matters of theory and outlook suggested by his paper. Calling himself a “behaviorist,” he summarizes the doctrines he imbibed around the lunch table in the Department of Psychology at Ohio State, doctrines about how through “language-reactions” to stimuli “the social group … becomes … a new and more complex type of organism.” “All this I have from Weiss, our psychologist,” he reports to Sapir – who must indeed have begun to wonder what this would-be chum was rambling on about with such puerile nonsense such as that he had already slammed down in his 1917 American Anthroplogist reply to A. L. Kroeber’s “The Superorganic.”

But Bloomfield’s discussion here is most revealing of how he, and, recapitulating him, Sapir, represented the achievements of “scientific” linguistics in the mid-1920s. Bloomfield goes on in his letter to explain why Weiss’s behavioristic psychology so appealed to him:

The first appeal this made to me was due to the fact that in our actual work we do exclude ‘mental’ factors, i.e., sound-change won’t stop for desires or fears. (Some of the Germans, I see from the Streitberg volume, are going in for the theory that ‘useless’ sound–groups are ‘more easily lost[’] – which seems to me to be pure bunk, resting upon vague ideas of what ‘usefulness’ and ‘meaning’ etc. are. Jespersen started it, with the notion that E[nglish] case-endings dropped off because no longer significant. I take it that a ‘useless’ case-ending is just as much and just as little meaningful and necessary to the speakers as any other part of the vocal series.)

As can be seen, what is at issue is the exceptionlessness of sound-law, which, as Bloomfield was to say in a footnote in his 1925 article on Proto-Central Algonquian, “is either a universal trait of human speech or nothing at all, an error.” Again, in his 1926 supplementary paper he terms it “a tested hypothesis: in so far as one may speak of such a thing … a proved truth.” It is in this light that we must understand Bloomfield’s report that “[he] ha[s] cut the angry reference to Meillet out of [his] article [which he in fact did not, ultimately], – all the more reason, if your Athabascan forms will give added proof, for not re-stating the fundamental laws of our science [!] every time someone forgets them.”

And so one can see that the place at which these two great scholars found common ground was in the great Neogrammarian achievement of the earlier century. Bloomfield’s “Postulates” of 1926, were to re-state Saussure’s Cours in a behaviorist and especially operationalist idiom (see Percy Bridgman, Bloomfield’s Harvard College classmate, on operationalist philosophy of science). And, doing so, Bloomfield followed the order of presentation of Saussure’s editors as of Sapir’s book Language of 1921, in that the synchronic nature of language comes first, and then the phenomena of language change – sound law, analogic change, semantic change, borrowing, etc. – constitute the dénouement of the whole (as well in Bloomfield’s own 1933 book Language, written at the University of Chicago). For Bloomfield, sound law rests on the postulate that “phonemes change,” i.e., that no matter what other kinds of changes do or do not affect a particular lexical form, its phonological shape can change independent of all these other factors, because the plane of structural phonology is taken to be relatively autonomous of all other planes of structure and function in language.

But more to the point: in 1927, by which time both Bloomfield and Sapir were at Chicago, an extraordinary paper appeared by the then young Jerzy Kuryłowicz (1895-1978), in which was emphasized not just the regularity of the assumption that “phonemes change,” but indeed the predictive value of the assumption of the regularity of sound change as an autonomous plane of diachrony. Saussure himself, in a burst of youthful genius, had done a reconstructive structural analysis of the phonology of the Proto-Indo-European lexical root, stem, and inflectional theme, and had postulated the existence of certain “coéfficients sonantiques” – later to be termed laryngeal consonants – in the parent language that had disappeared in the course of language history, but had left traces in certain colorations and quantitative alterations of adjacent root and suffix vowels. Bedřich Hrozný (1879-1952) had published his book on the decipherment of cuneiform Hittite in 1917, and in his startling 1927 paper in the festschrift volume for Jan Rozwadowski, Kuryłowicz shows that Hittite graphic ĥ and ĥĥ appear as reflexes in several of the positions where Saussure’s laryngeal theory had reconstructed them for Proto-Indo-European forms, the segments lost in the reflexes in the other branches.

Bloomfield and Sapir obviously understood the importance of this confirmatory discovery of heretofore unknown Hittite forms that confirmed, in the main, the Saussurean reconstructions that the young master had based on a structural phonemic analysis of lexical form. By early 1928, Bloomfield, whose “On the sound system of Central Algonquian,” recall, had appeared in the second issue of Language in 1925, had himself done fieldwork – courtesy Sapir and the Canadian museum – on a dialect of Cree, Swampy Cree (Manitoba), different from the one he had earlier used in his reconstructions. In that earlier work, he had postulated the existence of a proto-language cluster *-çk– on the basis of one stem, meaning ‘it is red’, *meçkusiwa in Bloomfield’s reconstruction, that showed correspondences among the four daughter languages slightly divergent from all the other examples. Sure enough, this new dialect of Cree, Swampy Cree, shows a distinct reflex, –htk-, precisely where the specially reconstructed cluster predicts something different.

When the Social Science Research Council’s Committee on Scientific Method in the Social Sciences was putting together its Methods in Social Science: A Case Book (ed. Stuart A. Rice [Chicago: University of Chicago Press, 1931]) in 1927 and 1928, it turned among others to leading members of the Committee, among them Edward Sapir, for case studies illustrative of how particular social scientists had used particular methods to yield breakthrough results in one or another of the fields of endeavor. Sapir’s essay, written in December, 1928, and revised in February, 1929, is not, as one might have expected, based on any of the achievements of recent synchronic linguistic breakthroughs in which he and Bloomfield were leaders. To recall, Sapir’s formulation of what became phonemics, “Sound patterns in language,” was a landmark that had appeared in the very second issue of the journal Language in Spring, 1925. Bloomfield’s “A set of postulates for the science of language,” in the journal the very next year, set out the entire field as a conceptual architectonic of postulates, definitions, and corollaries, seeming to fashion a rigorous “science” ab initio. By contrast, Sapir’s piece in the Methods case book is entitled “The concept of phonetic law as tested in primitive languages by Leonard Bloomfield” (pp.297-306). It comprises one of three analyses in section V., “Interpretations of Change as a Developmental Stage,” where J. B. Bury’s historical analysis of the concept of ‘progress’ is analyzed, as well as A.L. Kroeber’s studies of cycles in women’s sartorial fashion. As the editor, Stuart Rice, explains the grouping of papers in his “Introduction” (pp.12-13), here

Change is viewed somewhat in the light of an unfolding within a particular cultural entity itself. The papers in this section will have particular interest to those who are concerned with the possibilities of prediction in social science. Anaylsis 21, for example, [Sapir’s of sound law,] describes a foreseen discovery of linguistic forms which recalls to mind the predictions by astronomers of the discovery of Neptune and the newer trans-Neptune planet [presumably, Pluto].

Note the image projected for historical linguistics, akin to astrophysics in its power of prediction.

Sapir, recapitulating Bloomfield’s two Algonquian papers as well as his own even earlier work on Athapascan, projects a powerful historical “science” that, valid in North America no less than in Eurasia, both contravenes Meillet and in a sense outdoes Kuryłowicz (who was later to be Sapir’s student at Yale) on the terrain of American Native languages. Such was the way the two great Chicago scholars joined forces at the cusp of The Depression years. Both Bloomfield and Sapir were instrumental in what became, retrospectively in the 1930s, the great achievement of synchronic structuralism, the discovery of phonological structure and the application of parallel approaches to morphological and syntactic structure. Yet for Bloomfield, certainly, as I have already quoted, Pāņini’s synchronic linguistics had already made that breakthrough.

For him, the breakthrough of modern linguistics was elevating the Neogrammarian method of residues involved in modifying the Grimm’s Law regularities with Verner’s Law and Grassmann’s Law to a postulate of exceptionless of autonomous phonological change. Sapir’s book Language of 1921 had already stated, in somewhat nontechnical form, that the locus of sound law is in an abstract configurational system of ideal, inner sounds, in effect an abstract pattern of phonemic categories, and had reiterated this as the dénouement of his famous 1925 paper on “Sound patterns.” The discovery of the phoneme, then, allows us to see both that phonological systems are the loci of the sound law component of diachrony, and to make predictions about what shapes daughter forms should take. Sapir uses the occasion of his methodological paper to show parallel examples of this from Bloomfield’s Algonquian work and from his own work on Athapascan, where the predicted form had appeared in his data of Summer Quarter, 1927, in Hupa, a language of California.

If the synchronic phoneme, as Bloomfield wrote in 1927, made a science of linguistics, its predictive power lay in the very Neogrammarian diachrony that Sapir and his students would do so much to marginalize in the era of linguistic theory to come.

How to cite this post

Silverstein, Michael. 2014. In Praise of “Exceptionless”: Linguistics Among the Human Sciences at Bloomfield and Sapir’s Chicago. History and Philosophy of the Language Sciences. https://hiphilangsci.net/2014/11/26/in-praise-of-exceptionless-linguistics-among-the-human-sciences-at-bloomfield-and-sapirs-chicago

Tagged with: , , ,
Posted in 20th century, America, History, Linguistics

Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out / Change )

Twitter picture

You are commenting using your Twitter account. Log Out / Change )

Facebook photo

You are commenting using your Facebook account. Log Out / Change )

Google+ photo

You are commenting using your Google+ account. Log Out / Change )

Connecting to %s

%d bloggers like this: