Of Minds and Language (2 page)

Massimo Piattelli-Palmarini, Pello Salaburu, and Juan Uriagereka

This whole enterprise grew from a delightful equivocation. Everyone involved assumed we would be learning from Noam Chomsky, while he told us he was looking forward to the encounter in order to learn from the participants. We are convinced that the reader will benefit from this equivocation. It is a tribute to Chomsky and the other protagonists of this rich exchange that the layout of, and spirited exchanges upon, multiple central topics are among the most genuinely interdisciplinary to be found anywhere in the literature. We like to think that readers with quite different disciplinary backgrounds (linguistics, psychology, biology, computer science, or physics) will enjoy at least some sections of this book. The organization into parts and sections has been conceived with a view to facilitating such selective access.

The present ordering does not always reflect the chronology of the conference, though the discussions following each presentation, after minimal editing, are all reported here in “real time.” Most of the originality and interest of this volume lies, we think, in these candid discussions, but the reader, depending on concrete interests, may decide to go past some of them and connect to the following sections. In fact, although we tried to organize matters proceeding from the more general to the more specific, it was inevitable that, in the ensuing deliberations, specific, and even sometimes technical, issues be brought to the fore also for quite general presentations.

The book is divided into four parts, almost in contrapuntal fashion. The
Overtures
jointly offer different, but complementary, introductions to the central theme of this volume: biological perspectives on language and related cognitive functions. These presentations are all non-technical and, we think, accessible to readers with different backgrounds. The second part,
On Language
, is a multi-faceted attempt to draw the frontiers of an approach to
language seen as a natural object and, therefore, to linguistics conceived as part of the natural sciences. The third part,
On Acquisition
, focuses on how it is possible for every normal child to converge so rapidly and so efficiently onto the specific language of the surrounding community. Like the final entries of a fugue, the explorations in part four (
Open Talks on Open Inquiries
) enter domains of research that are conversant with, but also attempt to go beyond, the present concerns of linguistic theory (ethics, aesthetics, individual differences, neural correlates of emotion and prosody, and more).

In his opening remarks, Chomsky retraces the essential history of the field of biolinguistics and leads us to the present panorama. The chapters that follow explore, from different angles, the present contours of a biology of language. This part could be characterized, paraphrasing a famous paper by W. S. McCulloch,
¹
as an attempt to answer the question: What is biology, that language may be part of it?

Starting from very general questions and the premise that the more is packed into the Broad Faculty of Language, the easier it is to understand the overall evolution of this faculty (including its “narrow” aspects), Cedric Boeckx attempts to decompose Merge into more basic operations. He concentrates on endocentric (multiply nested, of the same type) structures specific to language, and seeks to derive this property from elementary “grouping” and “copying” operations, which he speculates may have been recruited from other cognitive systems in animal cognition. This fits into François Jacob's and Steven Jay Gould's dictum that new structures in biology are a recombination of old processes that are put together in new fashion, that being the general origin of evolutionary novelty.

Marc Hauser emphasizes the importance of probing the boundaries of animal cognition through “spontaneous methods.” He insists that there is virtually no connection in animals between the sensorimotor output of signaling and the richness of their conceptual systems. In order to bridge this gap, subtle experiments have been carried out to reveal the representation of the singular–plural distinction in monkeys and in prelinguistic children. Hauser then expands the analysis to the mass/count distinction, where he ascertains a contrast between monkeys and infants. He concludes with a proposal for the relations between language and ontological commitments which is sensitive to that mass–count distinction, so that it manifests itself only in some languages.

Charles Randy Gallistel explains why a materialist conception of mind is compatible with the attribution of high-level abstractions even to birds and bees. Experiments on the mastery by jays of thousands of locations of different food caches show that it is based on their memory of what they had hidden where and when. Moreover, on the basis of data on caching while being watched by conspecifics and then re-caching when out of view, Gallistel concludes that nonverbal animals represent the likely intentions, and reason about the probable future actions, of others. The mastery of solar ephemeris in the foraging bees demonstrates the sophistication of the spatial reasoning that goes on in these miniature brains. Such abstractions are both primitive and foundational aspects of mentation that must have emerged early in evolutionary history.

Gabriel Dover introduces a dissenting opinion. In contrast with Chomsky's plea for focusing on optimal computation in language design, Dover is hesitant to embrace the idea of a “rational morphology” that countenances only a limited number of archetypal body-plans. Detailing some factors in the present picture of evolution and development (modularity, redundancy, genetic regulatory networks, turnover, and degeneracy) Dover insists on a distinction in biology between the micro-level of chemical bonds – where the laws of physics are dominant – and a “higher” level where variation and “interactive promiscuity” reign. His position is that development is a “highly personalized” set of operations from the early inception of the networks regulating gene expression through to the ever changing neuronal connections in the brain. Subjectivity is the name of the game at all levels, even though we are only mindful of it in the brain.

Donata Vercelli, in stark contrast with that view, develops her considerations starting with the characteristics of a biological trait L (thinly disguised as being language) and stresses the importance for L of the dimension of plasticity. She then offers a summary of the mechanisms of epigenetics (under intense scrutiny in biology proper in the last half decade), suggesting that they may have a pivotal role in language development and may have had it too in language evolution. Vercelli and Piattelli-Palmarini conclude by suggesting that parametric variation across languages may well represent a genetic mini-max optimal solution, between the extreme of encoding every aspect of language genetically (thereby minimizing learning) and the opposite extreme of leaving all aspects of language to be learned (thereby minimizing the genetic load).

A counterpoint to Dover's view is also presented by Christopher Cherniak, who discusses his idea of a “non-genomic nativism.” As a result of computer calculations (previously published in detail by Cherniak et al. 2004), the minimization of connection costs at various levels of nervous systems in vertebrates
and invertebrates – from the placement of the brain in the body down to the sub-cellular level of neuron arborizations – emerges as being innate, though not genome-dependent. Models that also cover the optimal design of the best commercial micro-chips show that such optimal design comes “for free,” directly from the laws of physics. Cherniak's “non-genomic nativism” stresses the continuity between this finding and Chomsky's strong minimalist hypothesis, according to which narrow syntax is like a snowflake, shaped by natural law.

Still in the same spirit of McCulloch's quote, the second part of this book could be characterized as an attempt to answer the symmetric question to the one posed above: What is language, that it may be part of biology? This general theme is developed in various ways here, even conflicting ones. It is perhaps useful to keep in mind that James Higginbotham will, at the end of the conference, acknowledge that he and Luigi Rizzi identify themselves as being, in some sense at least, abstract biologists – a characterization that probably fairly describes all the language experts presenting their views in this section. That said, it is only natural for “natural philosophers” to explore views like these, rationally disagreeing when the evidence is conflictive.

Wolfram Hinzen defends the radically minimalistic view that structural semantic conditions are satisfied in virtually tautological terms with regard to a corresponding syntax. From his perspective, in effect only syntax is a natural system reflecting Chomsky's familiar “three factor” considerations, and it is (hopefully) rich enough to provide the essential scaffolding for semantic structuring. In a nutshell, syntax creates its own ontologies by virtue of its core mechanisms, and such ontologies are not independently given in any sense; the issue is to then match such ontologies with those needed to conceptualize, at least in their bare essentials. As Hinzen explains, this thesis extends the idea that language – if analytical tools for its structure go minimally beyond mere bracketing – and basic mathematics are virtually isomorphic.

James Higginbotham explores two putative interfaces of the linguistic system: one between syntax and semantics, and one between the latter and the world. The first implies asking how much of compositionality (the meaning of a whole being a function of the meaning of its parts and their mode of composition) belongs to general features of computation, as opposed to anything specific to language. A central issue is to explain where compositionality breaks down and what differences between languages should be explained in terms of parameters at the syntax/semantics interface. The second interface
involves the relations of semantics to our systematic beliefs about the world: What causes us to think/speak in the specific modes we do – and is this state of affairs necessary?

Sentences are known to ubiquitously contain parts that are interpreted not where they are pronounced. Yet there are strict, partly language-specific, constraints on what is syntactically allowed to be thus “moved,” where and how. Movement to distant sentential locations takes place via successive local steps, called “cyclical.” In his contribution, Luigi Rizzi argues that certain conditions on syntactic “impenetrability” can be derived from “intervention” – that is, effects arising when “movement” of a given element takes place over another
of the same type
. Locality is then relativized to skipping over interveners of equal or higher featural richness, so that elements involving fewer features have more leeway: when not involving, say, question sites, merely topicalized constituents result in less specified interveners. Thus, in the end only elements with rich featural arrays are forced into taking cyclic steps to by-pass “minimality” effects.

Juan Uriagereka discusses so-called uninterpretable features (Case being a paradigmatic example), which pose a puzzle for a minimalist program understood as an optimal solution to interface conditions. Why are there, then, uninterpretable features in languages? His suggestion is that their presence relates to a “viral” take on morphology: that is, the view that displacement correlates with the elimination of morphological specifications that bear no interpretive import. This abstractly recalls the workings of the adaptive immune system, and represents a solution to the parsing puzzle posed by compressing complex recursive (thought) structures into simple linear (phonetic) manifestations: the intricate syntax resulting from excising the viral morphology constitutes an effective instantiation of corresponding nuanced semantic types.

Complementing these approaches with a search for brain correlates to language, Angela Friederici's proposal is that the capacity to process hierarchical structures depends on a brain region that is not fully developed in monkeys, and that the phylogenetically younger piece of cortex may be functionally relevant for the acquisition of complex Phrase Structure Grammars. The older cortex may be sufficient to process local dependencies, while the human ability to process hierarchical structures could be based on the fully developed, phylogenetically younger cortex (Broca's area). Similarities and differences with germane studies on humans in other laboratories and with analogous inquiries by Hauser and Fitch into the processing limitations of grammars in tamarin monkeys, as compared to humans, emerge in the important ensuing discussion.

In the
round table on language universals,
Cedric Boeckx invites us to reconsider historically the very idea of language universals, and challenges the
notion of parameters as theoretically relevant in a minimalist framework, where universal grammar (or at least narrow syntax) is supposed to be genuinely universal, and all parametric variation (or at least its “macro” version) is discharged onto the morpho-lexicon. Janet Dean Fodor declares herself not so much as a “discoverer” of universals, but a “consumer” thereof. Fodor conveys the idea of how hard it is to explain the child's actual acquisition of grammars, concretely how laborious the process of hypothesis-testing is in the abstract. She candidly declares herself to be “shopping for” hypotheses that can constrain the acquisition of grammars in real life, to avoid hosts of overgeneralizations that are possible on paper, but that no child ever makes. Lila Gleitman emphasizes the puzzle of the acquisition of the meaning of “simple” verbs like hug or give for ten-month-olds, which combines the “poverty of the stimulus” problem with its virtual opposite: the richness of the stimulus problem. How does a baby know enough to ignore irrelevant accessory objects or events in a scene? She stresses that a mosaic of conspiring cues – each of them inadequate or even obfuscating by itself – are exploited by babies to converge, almost errorlessly, on the lexicon of their native tongue. Finally, Luigi Rizzi retraces the transition from generalizations about particular grammars to the principles of UG and the notion of parameter. He reviews the recent history of Principles and Parameters, from the Extended Standard Theory to consequences ensuing from the current Cartographic Program.