The Interplay of Linguistics, Computation, and Funding in the Cold War Era

In this interview, we talk to Judy Kaplan about universals in American linguistics of the mid-20th century.

Download | Spotify | Apple Podcasts | YouTube

References for Episode 43

Emmon Bach & Robert T. Harms, Universals in Linguistic Theory (New York: Holt, Rinehart and Winston, 1968)
Noam Chomsky, Aspects of the Theory of Syntax (Cambridge, MA: MIT Press, 1965).
Jamie Cohen-Cole, The Open Mind: Cold War Politics and the Sciences of Human Nature (Chicago: University of Chicago Press, 2013).
Joseph Greenberg, “Some Universals of Grammar with Special Reference to the Order of Meaningful Elements,” in Idem (ed.), Universals of Language (Cambridge, MA: M.I.T. Press, 1963), 73-113.
______ “The Nature and Uses of Linguistic Typologies,” IJAL 23: 68-77.
Roman Jakobson, Kindersprache, Aphasie und allgemeine Lautgesetze. (Uppsala: Almquist & Wiksell, 1941).
Linguistic Society of America, Committee on Endangered Languages and their Preservation, “The Need for the Documentation of Linguistic Diversity,” June 1994. (https://old.linguisticsociety.org/sites/default/files/lsa-stmt-documentation-linguistic-diversity.pdf)
Janet Martin-Nielsen, “A Forgotten Social Science? Creating a Place for Linguistics in the Historical Dialogue,” Journal of the History of the Behavioral Sciences 47(2): 147-172.

Transcript by Luca Dinu

JMc: Hi, I’m James McElvenny, [00:09] and you’re listening to the History and Philosophy of the Language Sciences podcast, [00:13] online at hiphilangsci.net. [00:16] There you can find links and references to all the literature we discuss. [00:20] Today, we’re talking to Judy Kaplan, who’s a historian of the human sciences and a Curatorial Fellow at the Science History Institute in Philadelphia. [00:30] One of Judy’s current projects is to investigate the notion of universals in American linguistics of the mid-20th century. [00:39] Why is it that various competing schools of American linguists in this period converged [00:44] on universals as the target of their research, despite their fundamental differences in scientific outlook? [00:50] What did they mean by “universals”, and what role did universals serve in their respective theories? [00:57] So, Judy, can you sketch the scene for us? [01:01] What was happening in American linguistics in the mid-20th century? [01:04] Who were the leading figures, and what positions did they take? [01:09]

JK: Thanks so much. [01:10] Viewed from above, we can see that the mid-20th century was a time of remarkable growth and expansion in American linguistics, [01:18] so funding went up, the first university departments were established, and the ranks generally grew. [01:25] These markers corresponded to a handful of different programs or positions. [01:30] As far as those leading figures and positions, most people concentrate on two dominant approaches to the study of language at this time: structuralism and generative grammar. [01:42] I should say that both were internally heterogeneous. [01:46] On the structuralist side, there were students of Edward Sapir in the anthropological tradition and Leonard Bloomfield on the more strictly linguistic side. [01:54] Structuralists focused on speech, a sort of self-conscious departure from 19th-century philology, on form, [02:01] on abstractions — ultimately, as the name suggests, on structure. [02:06] In terms of evidence, they were heavily invested in the study of Native American languages. [02:12] By the post-war period, the so-called neo-Bloomfieldians or distributionalists were at centre stage. [02:19] So, figures here are people like Charles Hockett, Zellig Harris, Bernard Bloch, Martin Joos, and others. [02:26] To your question, their position was basically that linguists should be looking at how linguistic forms show up in different speech environments. [02:33] They insisted on the distinction of different levels of analysis — so, phonology, morphology, syntax, these were all different levels. [02:40] And these points of emphasis pretty much bracketed meaning and the mind, which is important. [02:47] Much of the work was really mathematical in terms of its look and feel. [02:51] Emerging in a sense from this tradition, but also deeply critical of it, was Noam Chomsky and the study of transformational grammar. [02:58] This program elaborated transformational rules, and that was a term that was adapted from Harris, so you can see that there’s this kind of emergence. [03:06] So, it elaborated these transformational rules to link fundamental deep structures of language to surface manifestations, allowing linguists to turn from empirical data toward the study of grammar and the mind. [03:19] The emphasis was on theory construction, evidence of linguistic competence, and the definition of rules. [03:25] Though this program subsequently splintered into groups dedicated primarily to syntax and semantics, it did come to define a new mainstream. [03:35]

JMc: So, how do universals fit into this picture? [03:39]

JK: In a very broad sense, I think universals trace back to the supposed liberation of linguistics from philology — the question being, was linguistics going to be about language or languages? [03:49] So, if the former, then one might reasonably be on the lookout, at the end of the day, for features or phenomena that are universal in scope. [03:59] Just circling back then to the general lay of the land, it’s interesting that historians of science — so I come out of history of science — that they have tended to write about this with different points of emphasis, as opposed to linguistic historiographers or historians of linguistics, first and foremost. [04:16] Historians of science have characterized the relationship between these two approaches (the structuralist and the generative) as the contrast between behaviourism and mentalism, which was given all sorts of ideological overtones in the post-war period. [04:31] Linguists writing about their own history have wrestled with whether or not the remarkably swift emergence of transformational grammar constituted a scientific revolution in the Kuhnian sense. [04:42] It was unquestionably a remarkable turn, so participant accounts, institutional changes, interdisciplinary engagements, and citation patterns all make that really clear, [04:52] but there were also many linguists who were already trained and working in the structuralist tradition who felt that transformational grammar was not really for them. [05:01] And I might locate Joseph Greenberg somewhere in there, right? [05:05] So he was coming out of anthropology, first and foremost, he received his PhD in anthropology, and became really well-known by the 1950s for working on the genetic classification of African languages. [05:19] He was quite well regarded for that work. [05:22] He first turned to the study of African languages because he had been interested in Arabic and Semitic and had been approaching these from an anthropological standpoint, [05:31] and in 1957, he started to write about typology, and typology in connection to universals. [05:38] So Greenberg is taking a typological approach to the study of language universals, and he is leaning pretty heavily on the work of Roman Jakobson, who had introduced them in connection to the acquisition of child language. [05:52] One of the things that he says here is that it’s really important to consider implicational universals, so of the form “if X, then Y,” [05:59] and he sets out on a kind of data-gathering mission with this in mind. [06:04] And, you know, when he tees up the program — this was launched in a 1961 conference at Dobbs Ferry — [06:11] when he tees up the program, he says, you know, We have accumulated so much data at this point in time that the time is ripe to turn to the study of language universals. [06:22] Coming back to the question about how universals fit into the general topography or the landscape of post-World War II American linguistics, historians of the human sciences have shown how, in the wake of World War II [06:38] — which, you know, perpetrated unprecedented violence based on these racist and eugenic ideas that some human groups were distinct and less valuable than others — [06:47] liberal biologists and anthropologists, also psychologists, also I would add linguists here, came together to try to define a rosier and more inclusive picture of human nature. [06:59] So this was a picture that celebrated our best features as humans, our intelligence, our technological prowess, our cooperation, and also our common inheritance. [07:11] So I think that this is really important for understanding why this turned to universals in the mid-20th century. [07:17] That said, the way that Greenberg and Chomsky defined and operationalized universals research could not have been more different. [07:26] So Greenberg edged up to the study of universals defined in this typological way in a 1957 paper for the International Journal of American Linguistics. [07:36] This previewed his 1961 contribution to the Conference on Language Universals at Dobbs Ferry, “Some universals of grammar with particular reference to the order of meaningful events.” [07:46] This became known as the word order paper. [07:49] It’s really important and sort of set the agenda for this approach to the study of universals. [07:55] It gives a clear sense of what this research program was going to be all about. [07:59] It talked about things like sampling procedures, empirical evidence, the typology of word order of course, and it also offered some thoughts on the relationship between linguistics and other human sciences. [08:11] It was guided by the idea that the best, and you could say really the only, way to study universals was inductively. [08:19] So I mentioned before a debt that Greenberg had to Roman Jakobson, this idea of implicational universals, which appears I think in the first paragraph of that paper. [08:29] And so this, again, is this idea of “If you have this feature, you’re also going to have this feature,” or, you know, there’s going to be some other kind of constraint. [08:36] And this is just to point out maybe something that people haven’t necessarily thought about. [08:40] This is really a different way of saying what is commonly held as opposed to a statement like “All languages have Z.” [08:49] I think it’s also important to say something about the social organization of Greenberg’s research program. [08:56] As this was an effort to coordinate empirical generalizations on a massive scale, an army of researchers was needed. [09:03] They also needed to think carefully about how to manage their data, [09:06] so he was invested in the creation of some sort of cross-cultural file, he called it. [09:12] Greenberg said that coordinated efforts beyond the scope of individual researchers were going to be necessary to establish on firm grounds the actual facts concerning universals and language. [09:23]

JMc: So do you think this inductive approach of Greenberg is an example of this contrast that linguists often make between, you know, empiricism and rationalism in their approach to linguistics? [09:34] I mean, the generativists love to talk about that. [09:37]

JK: Yes, I think it’s important to differentiate between empiricism as a theory of mind and empirical research methods. [09:45] Mostly what I’ve been talking about here are empirical research methods and rationalism versus a sort of logical way of going about doing things. [09:54] So just to differentiate between the theory of mind and the methodology is important. [10:00] But yes, these do sort of map onto those basic traditions. [10:05]

JMc: Do you think it’d be fair to say that it’s part of generativist propaganda to conflate these two things and to set up this opposition? [10:13]

JK: Yeah, probably. [laughs] [10:15] I mean, I think the rationalist, in the sense of a theory of mind, description, applies to Chomsky. [10:22] It’s just that the empiricism, empirical method doesn’t necessarily apply to Greenberg. [10:28] But if you are in the generativist camp and you’re looking at what the linguists who are using empirical methods are trying to do, yeah, to sort of suggest that conflation does important work, [10:40] because it maybe consigns everyone who’s looking at actual attested language data to maybe this behaviourist tradition, which by the mid-20th century was sort of much maligned. [10:52]

JMc: How does Chomsky then fit into this picture? [10:53] So you’ve mentioned Greenberg. So did Greenberg start talking about universals and Chomsky sort of tried to catch up to him? [11:01]

JK: They both arguably were publishing on universals in 1957, [11:06] so there’s a real kind of direct coincidence here. The- [11:10]

JMc: And they both use the term “universals.” [11:12]

JK: Not exactly. So Chomsky uses it a little bit later. But the universals that Chomsky and his colleagues were interested in, [11:19] so this is on the other hand, they didn’t pertain to languages so much as they had to do with those internal aspects of linguistic theory that might be then regarded as universal [11:31] — so, for instance, the very idea that transformational rules can get you from meaning to sound via the grammar. [11:39] If these could be established, then these would call for some kind of further explanation, and this ends up being a claim about innateness, which is sort of characterized as a biological endowment. [11:51] Chomsky started talking about universals in an explicit way, to your question, in Aspects of the Theory of Syntax. [11:58] Here he links linguistic universals to the idea of explanatory adequacy and draws a really bright line between formal universals, that’s what he’s going to be interested in, and substantive universals, and this is more of the Jakobson-Greenberg cast. [12:14]

JMc: Okay, so does “formal universals” mean universals pertaining to theory construction about what linguistic grammars look like, whereas substantive universals are empirical findings about the structures of actual languages? [12:30]

JK: Exactly. He says early on in that text that “the main task of linguistic theory must be to develop an account of linguistic universals”, so this is a pretty clear statement, [12:41] and he urges that this account needs to be able to stand up to the actual diversity of human languages, while at the same time also accounting for the rapidity and the uniformity of language learning in rich and explicit ways. [12:53] So here he’s gesturing again to that kind of species-level endowment. [12:59] He gives examples of reducing or absorbing, I think he uses the word “abstracting”, phonological rules in English. [13:08] So here’s a, you know, what might be considered a substantive universal in the way that you just described, to the formal universal that a transformational cycle universally links phonological rules to syntactic structure. [13:22] And then this would have the advantage of explaining away funny-looking exceptions in terms of deeper underlying regularities. [13:31] If you look at the papers that were presented at the conference that was held on universals of linguistic theory… [13:36] So I’ve been thinking about these camps in terms of two organizing conferences. [13:40] So there was the one in Dobbs Ferry in 1961, and this other conference that was held, actually weirdly, on the same day, but just a few years later. [13:50] It was held, I should say, at UT Austin. [13:53] The kinds of universals, just to give a feel again, the kinds of universals that were up for discussion were very general things like case and modality. [14:03] They were derived primarily from common-sense intuition. [14:06] So if you contrast that again with the kind of empirical way in which Greenberg was working, they were mostly offered up by native speakers and they were mostly having to do with English language usage. [14:16]

JMc: So common-sense intuition, you mean these intuitions about grammaticality judgments? [14:21]

JK: Exactly. [14:22]

JMc: So whether you can say this or not, and put a star in front of it if you can’t. [14:25]

JK: Exactly.

JMc: Yeah, OK. [14:27]

JK: Yeah, and just to give one example of that, in his contribution to the conference, Paul Kiparsky suggested that rule addition and rule simplification might be called or considered universal processes of human language. [14:39] So you can get a sense of the feel that these are very, very different approaches. [14:42]

JMc: Yeah. OK. [14:44]

JK: The introduction to the proceedings that was published from that conference mentions that it was all tape recorded, and I’ve been trying to find the tapes [laughs] because, you know, yeah, the generativists are known for having these really rowdy conferences, [14:59] and I really want to hear what was actually said, but I haven’t been able to find them. [15:02] So if anybody listening knows where to find those tapes, please let me know. [15:06]

JMc: OK. So these two conferences, Dobbs Ferry and then UT Austin, who were the participants in these conferences? [15:15] So like, what was the sort of character of them? [15:18] So was the Dobbs Ferry conference more anthropological in orientation? [15:22]

JK: Yeah, so the Dobbs Ferry conference was funded by the SSRC, so the Social Science Research Council. [15:28] It grew out of conversations that were happening during… I want to say the academic year of 1958–59 at Stanford. [15:35] And, you know, as it’s described in the notes, there’s this sense that psychologists, sociologists, anthropologists, and linguists all needed to be talking about kind of universal phenomena. [15:48] So in terms of the participants at Dobbs Ferry, it was a very interdisciplinary group. [15:53] I think Greenberg says at one point that if we edge up to the universal, we’re going to go over this tipping point into psychology and anthropology. [16:02] As for UT Austin, Emmon Bach was there, Paul Kiparsky, I’m going to forget other names. [16:08] But just, you know, this was largely drawing on a pretty intimate group of actors in the transformational-generative school. [16:16]

JMc: And Greenberg’s statement about if we, you know, keep pursuing universals, we’ll end up doing psychology, is that intended as a reductive statement? [16:24]

JK: I don’t think so. [16:26] You know, again, he was really keen to work in an inductive way, so to amass a long list of generalizations, to take a step back from those to maybe elaborate even deeper fundamental principles, at which point you would be thinking you would have to confront psychology, [16:43] I think that’s more the spirit of in which he’s talking about this interdisciplinary engagement. [16:47]

JMc: You’ve sort of already touched on this idea when you mentioned that there was a movement among liberal scientists in America after the war to try and claw back some notion of our common humanity after the horrors of World War II, [16:59] but could you say a little bit more about why this idea of universals and this term “universal” was such a powerful concept for linguists? [17:08] And this was true of the broader social sciences, I take it, and even biology and medicine? [17:16]

JK: Yeah, well, focusing on linguistics for a second, I see the move towards universals, and both empirical and formal here, as an expression of sort of extraordinary disciplinary self-confidence in the middle of the 20th century. [17:27] So I talked at the beginning about how the discipline was on the rise, basically, and I think you have to be in that sort of powerful position to be able to take on or attempt to tackle something as ambitious as the universal. [17:39] It reminds me actually of the way that professional historians have sometimes put down antiquarian traditions. [17:45] So yeah, so just to say “We’re not going to look at these narrow phenomena; we’re going to look at something really ambitious and broad.” [17:52] In the writings of both Greenberg and Chomsky, you get the sense that describing the idiosyncrasies of language can be interesting, but that this is somehow too quaint. [18:01] So to come back to this idea of self-confidence, Greenberg says in his introduction to the Dobbs Ferry volume, and I’m paraphrasing here, that synchronic and diachronic linguistics has achieved sort of a high level of methodological sophistication, [18:15] and that also — and I think this is really important — that it has amassed a huge amount of data, such that the time was going to be really ripe for generalizing on a wide scale. [18:26] Chomsky says something similar, which is that traditional and structuralist grammars have provided extensive lists of exceptions and irregularities, but for him that meant that it was time to start moving beyond these to talk about underlying regularities. [18:40]

JMc: So how does this notion of universals — and the term “universal”, you know, which as you have just demonstrated, is actually being used in very different ways by people like Chomsky and Greenberg — how does this fit into what the broader human sciences were doing? [18:56]

JK: Other human sciences were talking about human nature, so I think that that’s where the point of connection is. [19:01] I haven’t necessarily seen universals as a term of art in other disciplines, but I think it’s really important to note — just with respect to this relationship between linguistics and the other social sciences, where it fit into the picture — it’s really important to note that linguistics was very successful. [19:19] It was really attractive to the folks, say, at the NSF who were holding the purse strings after World War II. [19:26] So the National Science Foundation, it might be helpful to clarify here, it made awards to both Chomsky and Greenberg. [19:34] So clearly this term, “universals”, was attractive to folks who had control of the money. [19:43]

JMc: So why did the NSF like linguistics so much? Was it because linguistics has the appearance, you know, with its sort of mathematization of grammar, has the appearance of a hard science, like of a real natural science, or could it be something else? [19:58]

JK: [laughs] That’s very intriguing. [20:00]

JMc: [laughs] [20:01]

JK: So… Right, so when the National Science Foundation was first established, it was not going to touch the human sciences. [20:08] It was not going to touch the social sciences at all. [20:11] It was established specifically for the physical sciences and engineering disciplines like this, [20:15] and so that’s how linguistics kind of snuck in, and with it, all of the social sciences. [20:22] So the first award that was ever made by the NSF to a project in the social sciences went to Chomsky, in fact, and the Research Laboratory of Electronics at MIT. [20:33] This was in 1956. [20:35] And, you know, it came as part of an initiative to support computation centres and initiatives and research having to do with numerical analysis. [20:45]

JMc: But did the NSF describe Chomsky’s project, like did they categorize Chomsky’s project as social sciences or as engineering? [20:53] Because, I mean, it’s in the research lab for electronics. [20:56]

JK: Yeah, as a computational initiative. [20:58]

JMc: Yeah. OK, so the NSF didn’t think of it as social sciences at all. [21:02]

JK: No, but it set up this precedent. [21:04] So the NSF was interested in funding linguistics because it was associated with computation and mathematics, and, you know, this set up a precedent where an emphasis on formalism and connections to engineering and the natural sciences — also, machine translation — were really, really at the top of the list of priorities. [21:25]

JMc: OK. So this sort of actually ties into a recent interview that we had with Chris Knight, you know, and Chris Knight, of course, has formulated the thesis that Chomsky, you know, deliberately made his work in generative grammar so abstruse so that it couldn’t be used in practice. [21:43] But he was being… Chomsky was being funded precisely because it could have practical engineering applications. [21:50]

JK: Yeah. Computer scientists saw a great deal of promise in what Chomsky had to offer, but he pretty quickly turned against computational projects, which is interesting. [22:03]

JMc: So, I mean, there’s this mathematization and possible applicability of linguistics [22:08] or possible engineering applications of linguistics, but are there any other features of linguistics as a discipline, as a social science discipline, that made it attractive to funders? [22:19] Perhaps the fact that it was, you know, so abstract and, you know, and mechanical? [22:26]

JK: Yeah, I think that’s really important. [22:29] As these federal funding agencies for scientific research were first getting set up, there was a desire to not be too overtly political, so the very, very abstract nature of linguistic research, and particularly, I would say, research on language universals, it fit the bill, [22:47] whereas projects in sociology or anthropology could have direct relevance to what was going on in the Cold War, say. [22:57] So there is something that was highly attractive about the really abstract nature of linguistic research. [23:03] I think… I’ve looked at correspondence from the Ford Foundation, and, you know, you have these sort of these program officers who are writing back and forth, trying to explain to each other what the linguists are doing, and they have a really tough job of it, and I think that’s actually really productive [laughs] for the field. [23:21] Another thing to say here is that in 1958, there was a really important piece of legislation that was passed, the National Defense Education Act, [23:29] and this is all, you know, sort of the history of big science. [23:34] And historians of science have really emphasized the way in which this led to, you know, expanded educational initiatives and funding for science and technology. [23:45] The piece that is often left out of that conversation is that it also sent a lot of money toward linguistics and linguistic training. [23:53] So the idea was that Americans had not bothered to learn other languages, and in the effort to sort of persuade international actors to join up with the American cause, that it’d be really, really helpful to actually be able to communicate. [24:13] And, you know, that might seem like a very practical goal, but there was a lot of research into how people actually learn languages and how this can be facilitated that was very abstract and basic and fundamental in nature. [24:26]

JMc: Was this discussion of universals a contained episode in the history of linguistics? [24:32] Did this universals discourse die down, or has it morphed into something else in contemporary linguistics? [24:39]

JK: I don’t know about contained, but I do think that the 1960s and early 1970s were a high watermark in the study of language universals. [24:48] You can see this… You know, so basically the argument that I’m trying to make with this research project is that if we think of linguistics as centrally having to do with this tension between the diversity of human languages and language as a species-defining function, or faculty I should say, those things are always in tension, but at the same time, when you look across time, I think that they sort of ebb and flow in terms of priority and succession. [25:14] You know, early 20th century before World War II, say, in American linguistics, this is a period that’s really defined by particularist interests in language. [25:23] By the time you get to Greenberg and Chomsky, it’s very, to use the word “universal”, universalist, [25:31] and, you know, following that time, there’s a return to a primary commitment to the particulars of human language, [25:39] and, you know, this gets bound up with concerns about linguistic diversity and the documentation of linguistic diversity. [25:47] One of my favourite sources on this is the 1994 statement on “The Need for the Documentation of Linguistic Diversity” that was put out by the Linguistic Society of America, and there, you could see that they’re looking back on this period in the 1960s and ’70s, and, you know, maybe even edging into the ’80s to some degree, although it gets messy, where typological research and generative research are really, you know, at the top of the discipline’s priorities. [26:15] And I think that that’s really interesting. [26:17] And so in this statement, you know, there’s this sense that these kinds of research programs (so the typological, the generative) need to somehow be folded into and inform a return to a primary interest or commitment to linguistic diversity. [26:35] So that tells me that this is a period that’s coming to an end. [26:38] It was unique. [26:39] It was bounded in some sense. [26:41] But yeah, that nevertheless, these were really, really important projects at the time. [26:46]

JMc: Well, thanks very much for answering those questions. [26:49]

JK: Sure.