Workshop #1

Human language in evolution: some key perspectives
organized by Shigeru Miyagawa (Massachusetts Institute of Technology)

Wednesday, 17 July 2019

Speakers
Koji Fujita, Kyoto University
Shigeru Miyagawa, MIT
Ljiljana Progovac, Wayne State University
Philippe Schlenker, Ecole Normale Supérieure Paris
Ian Tattersall, American Museum of Natural History
Emiliano Zaccarella, Max Planck Institute for Human Cognitive and Brain Sciences

Program

9:30 - 9:40	Introduction Shigeru Miyagawa This presentation provides arguments for a gradual emergence of syntax in human evolution, including hierarchical structure and transitivity, as subject not only to cultural innovation, but also to natural/sexual selection forces. I present a precise syntactic reconstruction of the ancestral, proto-grammar stage, characterized as an intransitive, flat, two-slot mold, unable to distinguish subjects from objects. Even this crude grammar offers clear and substantial communicative benefits over no grammar at all, as well as reveals, through its limits, reasons and rationale for evolving more complex grammars. The particular uses to which this protogrammar can be put even today (e.g. insult: cry-baby, kill-joy, tattle-tale; naming: rattle-snake; tumble-dung; proverbial wisdoms: Monkey see, monkey do; You sow, you reap) reveal why this cultural invention of coining binary compositions would have been highly adaptive at the dawn of language. By identifying insult (verbal aggression) as relevant for initial stages of language evolution, this proposal provides a specific sexual/natural selection scenario, which meaningfully interacts with the recent proposals invoking human self-domestication, especially as it relates to the reduction in physical aggression. Dealing with tangible and specific linguistic postulates, this approach lends itself to empirical testing through e.g. fMRI experiments, some results of which will be reported.
9:40 - 10:40	What use is half a sentence? Grammar caught in the act of natural/sexual selection Ljiljana Progovac, Wayne State University
10:40 - 11:40	The Evolution of Meaning: Problems and Prospects Philippe Schlenker, Institut Jean-Nicod and New York University
11:40 - 12:00	Break
12:00 - 1:00	Systems Underlying Human and Nonhuman Primate Communication: One, Two, or Infinite Shigeru Miyagawa, MIT (joint work with Esther Clarke, Durham U./MIT)
1:00 - 2:30	Lunch
2:30 - 3:30	Where did Merge Come From? A Sensorimotor Basis of Hierarchical Structure Building Koji Fujita, Kyoto University
3:30 - 4:30	Neuroanatomy of the Merging Mechanism in Humans: Topographical Organization, Ontogeny and Phylogeny Emiliano Zaccarella, Max Planck Institute for Human Cognitive and Brain Sciences
4:30 - 4:50	Break
4:50 - 5:50	The Pensive Primate:Language and the Emergence of Human Cognition Ian Tattersall, American Museum of Natural History
5:50	End of Workshop

Abstracts
What use is half a sentence? Grammar caught in the act of natural/sexual selection
Ljiljana Progovac, Wayne State University

This presentation provides arguments for a gradual emergence of syntax in human evolution, including hierarchical structure and transitivity, as subject not only to cultural innovation, but also to natural/sexual selection forces. I present a precise syntactic reconstruction of the ancestral, proto-grammar stage, characterized as an intransitive, flat, two-slot mold, unable to distinguish subjects from objects. Even this crude grammar offers clear and substantial communicative benefits over no grammar at all, as well as reveals, through its limits, reasons and rationale for evolving more complex grammars. The particular uses to which this protogrammar can be put even today (e.g. insult: cry-baby, kill-joy, tattle-tale; naming: rattle-snake; tumble-dung; proverbial wisdoms: Monkey see, monkey do; You sow, you reap) reveal why this cultural invention of coining binary compositions would have been highly adaptive at the dawn of language. By identifying insult (verbal aggression) as relevant for initial stages of language evolution, this proposal provides a specific sexual/natural selection scenario, which meaningfully interacts with the recent proposals invoking human self-domestication, especially as it relates to the reduction in physical aggression. Dealing with tangible and specific linguistic postulates, this approach lends itself to empirical testing through e.g. fMRI experiments, some results of which will be reported.

The Evolution of Meaning: Problems and Prospects
Philippe Schlenker, Institut Jean-Nicod and New York University

We argue that there are rich prospects for the study of the evolution of animal meanings, in particular in non-human primates. In a nutshell, comparative primate linguistics can be combined with standard animal phylogenies to reconstruct the evolutionary history of calls and gestures over millions of years, sometimes in a straightforward fashion. This contrasts with the enormous difficulty of reconstructing the evolutionary history of human language. One possible link between the two questions (still speculative at this point) pertains to the evolution of ape and human gestures, thanks to ground-breaking new work by Hobaiter and collaborators.

Systems Underlying Human and Nonhuman Primate Communication: One, Two, or Infinite
Shigeru Miyagawa, MIT, miyagawa@mit.edu (this is a joint work with Esther Clarke, Durham U./MIT)

Using artificially synthesized stimuli, previous research showed that cotton-top tamarin monkeys readily learn simple AB sequences, but not the more complex AnBn sequences that require hierarchical structure. Humans have no trouble learning both types of sequences (Fitch and Hauser 2004). Using similar artificially synthesized stimuli, Friederici et al. (2006) showed that there is a neuroanatomical difference in the areas of the brain recruited for the two types of sequences. While the simple AB sequence recruits the frontal operculum, the more complex AnBn sequence recruits not only the frontal operculum, but also the phylogenetically newer Broca's area (BA 44) (also Zaccarella and Friederici 2015). We analyzed reported vocal repertoires of Old World Monkeys to show that these nonhuman primates are capable of calls that consist not only of one item, but also calls that contain two items. Crucially, they cannot combine more than two items, as predicted from earlier research. We also show that the two-item calls cannot be the result of the combinatorial operation we see in humans, where the recursive application of Merge allows for a potentially infinite array of structures. The two-item calls are due to a dual-compartment frame (see Progovac 2015), which earlier research suggests resides in the frontal operculum, while the operation of Merge, which gives human language the potential for infinite array, resides in the Broca's area (BA 44). In this way, there is a sharp cut off after two: One, Two, or Infinite. We will also speculate on the emergence of the infinite system underlying human language based on our findings, drawing on earlier works by Stout (2008, 2010), Fujita (2017) and others.

Where did Merge Come From? A Sensorimotor Basis of Hierarchical Structure Building
Koji Fujita, Kyoto University

Human language is not a monolithic function but is an integration of several distinct systems each of which has an evolutionary history in other species independently of language. Among them, hierarchical syntax deserves a special attention as an apparent case of human autapomorphy, to be found nowhere in other animal communication systems. As such, it stands as a barrier to a natural understanding of human language evolution as it implies some evolutionary discontinuity between humans and other species. This barrier can be overcome once we note that similar hierarchical structures can be found in some non-communicative behaviors of other animals. In other words, we need to look beyond animal communication to fully understand how human language emerged.

In this talk, extending my previous proposals (Fujita 2017), I will suggest that the basic syntactic operation Merge evolved as an exaptation of the motor control capacity which manifests itself in the hierarchical and sequential object combination in tool use and too making. In her pioneering work on action grammar, Greenfield (1991) made a crucial distinction between Pot strategy and Subassembly strategy in combining more than two objects. This distinction can carry over easily to the two distinct recursive applications of Merge (Pot-Merge and Sub-Merge), the importance of which largely remains ignored or unnoticed in the generative literature.

On the observation that Sub-Merge is the core combinatorial device which makes human language possible at all, I will consider how the extension from 'action Merge' to linguistic Merge occurred only in the human lineage supported by metaphorical extension and multiple attention. The bottom line is that we need to bring generative grammar into closer contact with findings in other fields of research if we want to make it more relevant to studies of language evolution.

References:
Fujita, K. 2017. On the parallel evolution of syntax and lexicon: A Merge-only view. Journal of Neurolinguistics, 43B.
Greenfield, P.M. 1991. Language, tools, and brain: The ontogeny and phylogeny of hierarchically organized sequential behavior. Behavioral and Brain Sciences 14.

Neuroanatomy of the Merging Mechanism in Humans: Topographical Organization, Ontogeny and Phylogeny
Emiliano Zaccarella, Max Planck Institute for Human Cognitive and Brain Sciences

Humans understand infinite amounts of linguistic utterances in daily life. The syntactic structure of these utterances is thought to involve hierarchies of nested phrases, rather than linear sequences of individual words. According to a prominent hypothesis in theoretical linguistics, the human brain must be endowed with a simple combinatorial mechanism, conventionally called Merge, which drives linguistic comprehension to build hierarchies out of single words, on the basis of a finite set of syntactic rules. My goal is to characterize a plausible neuroanatomical hypothesis for such mechanism. In the first part, I will specifically exploit the possible internal subcomponents of Merge: (1) The stringing of elements in inputs; (2) The build-up of minimal hierarchies; (3) The recursive nature of such hierarchies; (4) The functional interface to the semantic system. I will then show preliminary experimental evidence on the neural causality, the directionality of information, the modality-independence (verbal, signed), and the domainspecificity (conceptual, motor) of the Merge subcomponents. By linking linguistic predictions to hemodynamic profiles, I will finally put forward a topographical neural model for Merge, which assigns peculiar structure-labeling roles to Broca's area, under the tenet that language faculty must consist of some neural computation capturing universal combinatorial power on one side, and particular categorical restrictions on the other. In the second part, I will first dig into the ontogenetic nature of Merge as reflected in the syntactic diversity measures of two-word combinations from extensive corpus collections of children's speech, and I will conclude the talk by considering possible phylogenetic trajectories linking humans to non-human primates.

The Pensive Primate: Language and the Emergence of Human Cognition
Ian Tattersall, American Museum of Natural History

Modern human beings process information symbolically, rearranging mental symbols according to rules to envision multiple potential realities. They also express the ideas thus formed using structured articulate language. No other living creature does either of these things, reflecting a qualitative cognitive gulf between modern Homo sapiens and, as far as we can tell, every other species in the Great Tree of Life that unites all living beings on the planet. Yet it is evident that we are descended from an ancestor that was both nonsymbolic and nonlinguistic. How did the astonishing transformation to modern cognition occur? Scrutiny of the fossil and archaeological records reveals that the transition to symbolic reasoning happened very late in hominid history - indeed, within the tenure of anatomically recognizable Homo sapiens. It was evidently not simply a passive result of the increase in brain size that typified multiple lineages of the genus Homo over the Pleistocene. I propose that a brain exaptively capable of complex symbolic manipulation and language acquisition was acquired as a byproduct of the major developmental reorganization that gave rise to the anatomically distinctive species Homo sapiens at about 200,000 years ago, and that the new symbolic capacity it embodied was recruited later, through the action of a behavioral stimulus. In evolutionary terms this would have been a rather routine event: after all, any structure must necessarily be in place before it can be used for a new purpose. Given the intimate interdependence of modern cognition and language - both are intrinsically symbolic activities - the most plausible cultural trigger for symbolic thought processes was the spontaneous invention of language in an African isolate of early Homo sapiens at (very approximately) 100,000 years ago. Language has several advantages in this role relative to other putative stimuli such as theory of mind.

Venue
Student Cultural Center Xenia (see map)
16, Sofokli Venizelou str
74100 Rethymnon Crete

Support for this workshop is provided by MEXT/JSPS Grant-in-Aid for Scientific Research on Innovative Areas 4903 (Evolinguistics) from the Japanese national government.

---