CARTA: Language Structure and the History/Future of Lingustics with Robert Kluender
We are the paradoxical ape. Bipedal, naked, large brain. Long the master of fire, tools, and language, but still trying to understand ourselves.
Aware that death is inevitable. Yet filled with optimism, we grow up slowly. We hand down knowledge. We empathize and deceive.
We shape the future from our shared understanding of the past. CARTA brings together experts from diverse disciplines to exchange insights on who we are and how we got here. An exploration made possible by the generosity of humans like you.
[MUSIC] Early in my career, I took a gamble on neurolinguistics as a research path before it was even really a thing. It was just dumb luck on my part, a case of stumbling onto the right thing at the right time. Now, neurolinguistics is firmly entrenched as its own sub-specialization within linguistics, complete with its own society, journal, and annual conference venue. In that sense, I do have something of a track record in anticipating future trends in the field, as evidenced by this piece I wrote around the turn of the century. On the other hand, for decades, I've been telling anyone who would listen to me, but I suspected that genetics or neurogenetics would be the next big thing in linguistics.
Well, that may still be the case, so far I can't really say that this prediction has panned out as well. This is not to say that there is no there, but so far the genetic component of language hasn't come into sharp focus. Even though we know full well that genetics certainly plays a role in language ability as recently reaffirmed by this study. What I'd like to do in this talk, is first look backwards in time to ponder where we've come from in order to get a sense for where we are today, and where we might be headed in future.
Since I doubt that many of you are familiar with the history of linguistics, let me give you a selective cursory and warp speed helicopter overview of it. Just about any intro to linguistics course will sooner or later point out that a lot of what we do in linguistics has its roots in what the Sanskrit grammarians of India did several millennia ago. However, their analysis of language was deeply rooted in the ritual culture and religious practices of the time. The primary and arguably sole aim of analyzing language was to preserve its efficacy in the performance of the ceremonial rights that it accompanied. Likewise, I don't think it's too much of an exaggeration or oversimplification to claim that linguistics in the modern era really got its start with the discovery, quote unquote, by European imperialists of Sanskrit, the mother language of many, but not all of the modern Indian languages.
Thanks in large part to the colonizing juggernaut of the British East India Company and it's private huge military forces. The co-founders of the Asiatic society of Bengal in 1784, which was modeled after the British Royal Society, were first of all, a typographer employee of the British East India Company, Charles Wilkins, and the British equivalent of a superior court judge in colonial India named William Jones. Both had studied Sanskrit with Indian pundits who incidentally were themselves barred from membership in the Asiatic society until 45 years after its founding. The discovery traditionally but erroneously attributed to Jones, of the distant relationship of Sanskrit to most of the European languages gave rise in the 19th century to the heyday of so-called philology.
The study of how languages are related to each other and develop historically, referred to in North America nowadays as historical linguistics. It's probably not entirely an accident of history, but this important period in the study of language ran parallel to that of biology and particularly to Darwinian theory of the 19th century. Virtually all of the 19th century linguistics superstars were a part of this philological tradition, including Jacob Grimm of Brothers Grimm fame, known in linguistics for Grimm's Law, which established these sound correspondences between Romance and Germanic languages. The bridge from the philological tradition of 19th century to 20th century linguistics was Ferdinand de Saussure who is famous for two things. First, having posited the linguistic equivalent of the Higgs boson before it was discovered in the development of the Indo-European languages, and second for laying out the structural principles of modern linguistics in a series of lectures at the University of Geneva, reconstructed from lecture notes by his students and published post humorously as course in General Linguistics. Serious structural principles of language spread throughout Europe, not only in linguistics, but also within the field of literary criticism.
Two of the three separate but related schools of European structuralist thought, namely those founded in Moscow and later in Prague, largely under the influence of Roman Jakobson, applied structuralist principles to both linguistics and literary criticism. Jakobson fled the Nazis and ended up in New York where he was a founding member of Lecole Lible Des Haustes Etudes, established for French-speaking scholars after the German invasion of France in 1940 as part of a new school to which incidentally, Parsons School of Design still belongs. It was here that Jakobson came into contact with and influenced Claude Levi-Strauss, the French anthropologist who adopted structuralist principles and famously applied them to the study of human culture, notably kinship terms and later myth, which is by the way, the topic of the spring CARTA symposium. In American structuralism interestingly, the influence tended to run in the opposite direction. Linguistics was heavily influenced by cultural anthropology instead.
This was promoted by the intense research effort to study native North America culture and language, as both became increasingly endangered in the late 19th and early 20th centuries. Franz Boas, often considered the father of the scientific study of language in the US, and his brilliant student Edward Sapir, both had life long appointments as anthropologists with research interests in language. However, Boas was primarily interested in language as a means to study culture. Sapir had extensive interests outside of linguistics and anthropology, notably in psychology and in the fine arts. The other giant of early 20th century linguistics, Leonard Bloomfield, was a different case altogether. Bloomfield was never an anthropologist, although he too did field work on many native North American languages and always held at least joined academic appointments in philology or linguistics.
Bloomfield champion the autonomy of linguistics as a field of study in itself, independent of anthropology and psychology. It was also Bloomfield who sent linguistics down the path of behaviorism. The Chomsky revolution of the 1950s overturned that entire paradigm and returned linguistics to its mentalist foundations. It preserved the autonomy of language from other fields of study that Bloomfield had fought for.
All of this has been requisite background for the possible path going forward that I'm going to lay out. But first, one more preliminary case study from the history of linguistics, namely the study of sign language. As many are now aware, American sign language was not even recognized as a linguistic system until the late 1960s and early 1970s.
Even though it had been in existence since the early 1800s. As a result, in the early years of sign language linguistic research, there was a concerted effort to distinguish the linguistic properties of sign language from mere gesture, as well as an emphasis on the arbitrary nature of linguistic sign, one of sociares principles of language. The form of a linguistic sign is generally taken to be independent of its meaning. With sign language, this is a trickier argument to make. As many, though not all linguistic signs have iconic origins. Meaning they do often in some way resemble the things that they represent.
However, this resemblance usually becomes attenuated and conventionalized over time, such that one often can't recognize the original iconic relationship without being told what it is. But now that sign language linguistics has staked out its conceptual territory, sign language linguists have begun to show renewed interest in investigating the iconic origins of many signs, as well as what is referred to as the gesture sign continuum namely, how gestural systems evolve to become fully linguistic. Now the argument I would like to make here is this, just as sign language has by now secured its status as a bonafide linguistic system and can therefore afford to examine more closely its own iconic and gestural origins without fear of compromise, so too can the field of linguistics as a whole and the study of language evolution in particular, afford to loosen the stringent separation from its cultural and psychological and even artistic roots imposed over the past century. It's certainly the case that as mentioned at the outset, neurolinguistics has in recent decades gained a foothold within linguistics and psycho-linguistics and sociolinguistics have been around even longer. But while these sub-disciplines are by now well established, they really bear only on the instantiation or implementation of language in real time and space. They aren't recognized as having much to contribute to the actual structural analysis of language.
What I'm going to do next is show you one area of research I've been involved in with my graduate students, Emily Davis over the past few years. That has led me to wonder if it might not be a good idea for linguistics to look a little farther afield than we've been used to and specifically to those companion areas of study from which linguistics arose and with which it was to varying degrees integrated throughout its history. If you don't pay close attention to debase within language evolution for which you can be forgiven, for the past 20 or 30 years, there's been something of an elephant in the room. That elephant is recursion, the embedding of similar structures and increasingly higher levels of analysis as shown in this example that you may be familiar with containing recursive relative clauses. In language, recursion generally looks like this. The host knew the man who brought the woman who left.
Which branches to the right for languages like English or to the left in languages like Japanese. But it can also look like this. An example of so-called center-embedded recursion. The woman, the man, the host knew, brought, left.
You may well think that's just not a grammatical construction in English, but it actually is. It means basically the same thing as the right branching version. The information is just packaged differently.
The grammatical rules of English freely and legitimately generate this structure, as can be shown by changing the nature of the subjects in the embedded clauses. The woman, someone I knew, brought, left. Now you should have no problem understanding this construction. Even though the structural properties of both sentences are exactly the same. Linguists have long recognized this fact. This early study by Miller and Isard in 1964 concluded that these structures present a processing problem tied to limitations of short-term memory rather than a grammatical problem.
Twenty years ago now, this science paper by Houser, Chomsky and Fitch threw down the gauntlet in claiming that with regard to the evolution of language, nothing matters other than recursion. In this paradigm, basic perceptual, motor and other cognitive abilities that support language, aside from recursion, were more or less shunted to the periphery and not unreasonably posited to be shared, at least in rudimentary form in common with non-human animals. The challenge that this approach poses is that it renders any plausible adaptive evolutionary account of recursion extremely difficult as there's precious little evidence of disability in the animal kingdom. But the authors of this paper simultaneously acknowledged that there could be another evolutionary path for deriving complex traits, such as recursion in language that can automatically be ruled out and that is acceptation. One picture is worth 1,000 words in this case.
The feathers have dinosaurs are presumed to have evolved originally for purposes of thermoregulation, but then over time began to be used as well for sexual display and were eventually accepted for flight, which drove their further evolution. How's their Chomsky and Fitch noted that recursion could have been excepted from other non-communicative domains of animal cognition, such as navigation, numerosity, or the structure of social relationships. This concession led to experimental investigations of the ability of non-human animals to produce or recognize recursive patterns and more specifically, center-embedded palindromic sequences. This study by Jiang et al tested the ability of macaque monkeys to reproduce a sequence of light flashes in either the original or reversed order, the latter of which is shown here. More impressively, Abe and Watanabe trained Bengalis finches to reliably associate pairs of notes extracted from their natural songs and center-embedded such pairs of nodes within each other. There are two things to note about these studies.
What is that? The form of recursion that the animals learned is center-embedded. The other is that clearly language is not a prerequisite for the ability to recognize or produce center-embedded structures because Bengalis, finches, and macaques can be claimed to do it and yet clearly don't have language themselves. This body of animal research was something of an aha moment for me because I've long been a fan of the work of the late Frits Staal, a Sanskrit scholar, linguist and philosopher of language for decades at UC Berkeley. In 1975, Staal helped to facilitate and organize a rare 12-day, 3,000 year-old Vedic right performed in the state of Kerala at the extreme Southwest tip of the Indian subcontinent. Staal was allowed to document this 1975 performance in detail.
This event made a huge impression on him and occupied his thinking for much of the rest of his career, he came to the same conclusion as other scholars of ritual. Namely, what was important about these rituals was their internal structure and not their purported meaning. In fact, styles analysis of Vedic ritual led him to conclude that more importantly, there was a form of hierarchical center-embedding built into its structure, as you can see Example 5 here.
This led him to the hypothesis that the use of center-embedding in human and by extension in hominin ritual practice and perhaps also in animal ritual, could have predated its use in language. In other words, center-embedding, a form of recursion could have been accepted for material practice. This was to put it mildly, a controversial hypothesis. The question that this hypothesis raises is this, if central-embedding and ritual practice underlies the evolution of recursion in human language, then why don't we see more evidence of it? Center-embedding is famously infrequent in human language for reasons we saw earlier, it's often really hard to process in real time. Staal observed that this was not a problem with regard to the structure of ritual, however. As it unfolds in a much more luxurious, leisurely timescale for days.
As it turns out, this is equally true of a number of art forms that trace their origins directly back to human ritual culture. Despite my lifelong interest in the arts, I was until recently completely unaware of the fact that principles of center embedding play an equally prominent role in narrative poetry, music, and architecture. The classic Indian epic, the Mahabharata, consists of multiple embeddings of stories within stories within stories or so-called frame structure. In fact, Minkowski suggested that this structure was borrowed from ritual. This is just one example of a common literary device and means of literary analysis called ring, or wheel symmetrical composition, which has been identified in books of the Hebrew Bible and in epics as culturally diverse as the Iliad and the Odyssey, the Aeneid, the song of Roland, Beowulf, 1,001 Arabian Nights [inaudible] and the Canterbury Tales.
As well as in Shakespeare's plays, Mary Shelley's Frankenstein and [inaudible] . However, center embedding is not merely a feature of entire works, but something that can be discerned at every level of literary analysis, for example, the so-called riddle hymns of the Rig Veda embed the enigma in the middle of the hymn, and bracket it off with parallel stylistic, lexical and morpho syntactic framing devices on either side of it in preceding and following versus indicated here in yellow and green. This symmetric form has been identified in episodes and verses of the epics as well. Center embedding is likewise ubiquitous in Western music composition in every period and at every level of analysis, using musical gestures such as themes, motifs, rhythm, texture, color, and harmonic progressions. Aside from the Sonata, symmetry is also built into the minuet and the roundup.
The arch form in which passages or movements are mirrored on the basis of key, tonal, center or contour. Retrograde composition in which the music is identical on some parameters such as rhythm or pitch, when played forward or backward, are common musical concepts. My graduate student Emily Davis noted that this type of symmetrical structuring extends to architecture of which one of the foundational principles is in fact, symmetry. This can be seen at the level of the most basic, traditionally temporal edifices constructed for Vedic rituals, subsequently burned to the ground at the conclusion of the rights to the later permanent edifices of Indian Hindu and Muslim culture.
Actually in architectural traditions across cultures. This symmetry likewise extends to many forms of artwork, ranging from ancient Greek pottery to contemporaneous configurations of images in today's digital world. With regard to the question, is central embedding is so fundamental to the use of recursion in human cognition.
Why don't we see more of it? The answer is that culturally speaking, its remnants are actually pretty ubiquitous and the domains in which we have thus far identified it, namely in storytelling, music, and visual art are not just random aspects of human cultural endeavor, but intrinsic aspects of human ritual. As any social or cultural anthropologist who studies ritual can tell you. This brings us back to Staals original proposal that the recursive structure of ritual ultimately provided the basis for the recursive structure of language. Coming full circle now, let me return to the main point of this talk. In taking a stroll down memory lane of the history of linguistics, I tried to emphasize the degree to which linguistics as a field and intimate connections with, as well as roots in related fields of scholarly endeavor, including cultural anthropology in literary studies. The legacy of Bloomfield in the 20th century has been the intellectual autonomy of linguistics as a field of study.
This independence of linguistics from its sister disciplines has resulted in amazing progress in the study of language over the past 65 years or so. However, the sands of time have been shifting in linguistics since the turn of the century and I think it's perhaps not unfair to ask what comes next at this point. As the dominant Chomsky and paradigm begins to lose some of its inherent force. My suggestion here is that just as the study of sign language has taken the brave step of re-evaluating its origins in gesture and iconicity, so too would it behooved the study of language more generally to cast its net a little wider than has been the case in recent practice. This is not a novel idea.
Studies in the evolution of language have for years been emphasizing the role that cultural factors play in shaping the structure of language as it is transmitted from generation to generation. All I'm suggesting here is that the more general study of language might benefit from casting it's glance backwards in time more often towards these more cultural influences. Also from entertaining the idea that enlisting the support if it's sister fields, especially anthropology, psychology, and the arts, along with neuroscience and genetics, may help to broaden its scope and range of application in a profitable way to everyone's general benefit. Thanks for your attention. [MUSIC]