Reflections on Syntax

Lectures in General Linguistics, Syntax, and Child Language Acquisition

by Joseph Galasso (Author)
Monographs XXXVI, 278 Pages

Table Of Content

  • Cover
  • Title
  • Copyright
  • About the author
  • About the book
  • This eBook can be cited
  • Table of Contents
  • List of Figures and Tables
  • Preface
  • Overview
  • Introduction
  • 1 Opening Philosophical Questions: Language and Brain Analogies
  • 2 Preliminary Overview
  • 3 The ‘Four Sentences’
  • 4 Reflections on Syntax
  • 5 Reasons for Syntactic Movement/‘Four Sentences’ Revisited
  • 6 The Myth of ‘Function Defines Form’ as the Null-Biological Adaptive Process and the Counter Linguistics-Based Response. (The ‘Accumulative Lecture’)
  • Appendixes
  • A1: Poverty of Stimulus
  • A2: Concluding Remarks. The Dual Mechanism: Studies on Language
  • A3: A Note on ‘Proto-language’: A Merge-Based Theory of Language Acquisition—Case, Agreement and Word Order Revisited
  • A4: Concluding Remarks: Lack of Recursion Found in Protolanguage
  • A5: A Note on the Dual Mechanism Model: Language Acquisition vs. Learning and the Bell-Shape Curve
  • A6: Overview of Chomsky
  • Works Cited
  • List of Terms (informal definitions)
  • Full References and Web Links
  • Index
  • Series index


Whether it be … constraints placed on phonological assimilation which stipulate that in order for the horizontal spreading of voicing to occur between two adjacent consonants, they must first be of a ‘sisterhood’ relation; whereby, for example, the /r/ in ‘cars’ provokes assimilation of plural /s/ => /z/, in contrast to the /r/ in ‘Carson’ which does not. (The former a structural ‘sisterhood’ relation, the latter a ‘mother-daughter’ relation in terms of ‘family-tree’ hierarchy):

Assimilation of ‘voicing’ applies between sisters.

*No assimilation between mother-daughter.

←xi | xii→

Or whether it be … the naïve view that the two apparently adjacent final-position sounds of /_ks/ as found in the two words ‘fix’ /fIks/ versus ‘speaks’ /spiks/ surely must equally get processed similarly (they do NOT) calls on us to reconsider something much more insidious going on in the underlying structure of morphophonology:

The differences in processing resulting in such distinctions can account for both developmental (as found in child language) as well as for Second Language (L2) errors of omission: viz., with a lexical stage-1 of child language often deleting the final ‘mother-daughter’ /_/s/ inflectional affix but never deleting the final ‘sisterhood’ /_s/ stem-based element (with similar findings for L2 adults).

And then … to think that such scaffolding of ‘mother-daughter’ hierarchical structure which yields a recursive syntax comes to us for ‘free’—part-and-parcel of the design of the human brain/mind—is something to wonder. This is what this book is about—’the wonder and unfolding of recursive syntax’, and the manner in which it has forced the field of modern-day linguistics to reconsider old assumptions we once held dear—old assumptions which were hard to kill off, but which had to eventually die at the stroke of the generative grammar enterprise.

The chapters contained in this book derive from a series of accumulative course lectures given across a span of several semesters to my graduate students of theoretical syntax, as well as to my many undergraduate students of child language acquisition, both at California State University Northridge, as well as Cal State Long Beach where I have lectured as an adjunct professor over the past twenty years. I’d like to thank all my students over the years that have helped shape these lectures. Our collective class discussions have better sharpened my own understanding of these issues. If these lectures in linguistics have improved at all since their first incarnation, it is only because they have benefited from the many discussions, multifaceted argumentation, and the steadfast persistence on seeking-out diverting points of departure on given topics—all respectively instigated by you, my students, over those years.←xii | xiii→

These lectures are immensely Chomskyan in spirit, recursive-syntactic in nature, and are tethered to a framework which takes as the null hypothesis the notion that language is an innate, pre-determined biological system—a system which, by definition, is multi-complex, human-specific, and analogous to a philosophy highly commensurate of Descartes’ great proverbial adage which announces the calling for a ‘Ghost-in-the-machine’.

And for those today who would wish-away Descartes’ Mind-body dualism as no longer tenable, Chomsky turns the table on them by suggesting that all we have achieved, thus far, is exorcise the machine (via Newtonian mechanics), we have left the ghost intact. So, if philosophical dualism is claimed to be no longer tenable (which I believed is a false claim), then it is not for the typical reasons assigned to the break. Rather, dispensing with a duality, all we are left with is the singular haunting ghost. (Chomsky 2002, p. 53).

Conceptually, the production of this book comes out of an abridged edition of my theoretical monograph entitled ‘Recursive syntax’ (LINCOM Publications, Studies in Theoretical Linguistics, 61, 2019) whereas a conceptual pedagogical device I address the syntactic implications of the Four Sentences. While the ‘four-sentences’ came to me merely as a point of departure, as a sort of omnibus tour of Chomskyan syntax over the last half of the past century, it also occurred to me to show how recursive designs of language—i.e., Reflections of Syntax—might play a significant role in so many different spin-off areas. These after-thoughts formulate much of the material found in the appendixes of the text.

I’d like to thank all my colleagues of the faculty of linguistics at California State University—Northridge where I have been a proud part of this fine theoretical department over the past twenty years.I’d very much like to thank all involved with the production of this text: Tony Mason, Jackie Pavlovic, Abdur Rawoof, as well as Naviya Palani along with all the editors on her production team at Peter Lang.

←xiii | xiv→


←xiv | xv→

A Brief History of Psychology. Let’s began with some interesting and historical analogies related to (i) the technology-interface to learning, and (ii) brain-analogies. It’s interesting to question what the many psychological impacts have been on the state of our human evolution. For instance, we can start with the invention of paper and what its lightweight, efficient economy and easy transport has meant for the establishment of learning. (I am reminded of the discovery of the Archimedes Palimpsest, the 10th century manuscript found when its original vellum, the dried animal skin used before paper, revealed what was just underneath its surface—as monks, three centuries later, recycled earlier vellums by scraping-off the prior script. We can only be horrified by the sheer volume of writing lost over time). Of course, the typical inventions follow—all of which bring very different psychological impacts: the (movable type) printing press and how the eventual spread of knowledge (sciences, religion) played on our human psychology. The typewriter, the PC computer, advancing software (the ability to cut & paste and copy), the floppy-disk … through to all the trappings of the ‘internet’ (first called the ‘ethernet’, and then the ‘information superhighway’: metaphors for ‘ethereal & otherness’—the neither ‘here nor there’—and of unfathomable ‘speed’). These innovations are often reduced and treated as ‘hardware’ developments, as artifacts—but it is indeed interesting to ask, in retrospect, what such ←xv | xvi→incremental progress meant for our human psychology, what it meant for our human, biologically-based ‘software’ (i.e., the human mind). It is instructive to look for psychological impacts and to ask how human experiences regarding our interface with such innovations have helped shape our understanding of ourselves, our fellow man, as well as the world around us.

Brain Analogies. This led to so-called (historic) ‘brain-analogies’. For instance, it was once understood that the brain was analogous to mechanical ‘clocks’ (of the philosophical ‘clouds & clocks’ argument, see Karl Popper ‘Objective Knowledge: an evolutionary approach’ Ch. 6). In this antiquated notion, the ‘brain as clock’ was said to be made-up of levers and gears which would interact in very trivial ways with the environment. The most obvious interaction with our brain-as-clock metaphor was to count and remember things. A person was understood to be the mere product of the things we came across in our environment, the things we noticed and counted: ‘man as ultimate calculator’. (See Locke’s notion of man as a blank slate, a tabula rasa). Whether or not a person was ‘smart’ was based on how well he noticed and remembered his token counts of environmental interactions; of course, there was no notion as to why a person might notice one thing over another (that question might have more to do with the psychology of observation, ‘a cloud’). The idea of how a ‘bad’ experience might impact our brain/clock—for example, how a person’s personality might be affected and formed—was not considered. In fact, such ideas of ‘personality’ really don’t begin to be formulated, psychologically, until the 19th century, coming to bear on the work of Freud, etc. (But there were earlier antecedents for sure, found in 17th century early modern English literature: e.g., Shakespeare’s first psychological profile of MacBeth).

The notion of ‘brain as cloud’ begins to take shape actually in the second half of the 20th century with the work of the linguist Noam Chomsky, who himself brings forth 17th century arguments from the Age of Enlightenment’s own René Descartes in forming a ‘Cartesian linguistics’. This new ‘spookiness’ of the brain (a nebulous-like cloud)—an enlightenment invention—was a large part of the underpinnings of Chomsky’s linguistics theoretical pursuits: viz., how Chomsky questions the ‘direct-environmental’ theories of language acquisition of the day. A clock account to language acquisition would have it that the child’s sole role in language acquisition is to consciously count the linguistic items she comes across in the course of her language growth. This direct ‘input-to-output’ imitation-scheme was a large attribute of the Behaviorists Theories behind general learning procedures, as advanced by the Behaviorists schools of thought, e.g., B.F. Skinner. (See imitation to analogy to computational below).←xvi | xvii→

This classic dichotomy over Empirical thought (e.g., Aristotle, John Locke, John Stuart Mill, and B.F. Skinner) versus Rational thought (e.g., Plato, Descartes, Spinoza, and Noam Chomsky) has remained constant over time, with only the refashioning of terminology making the epic debate seem contemporary. For instance, the current developmental/psychology debate over ‘Nature vs. Nurture’, more currently coined as ‘the nature of Nature’ is nothing more than a reworking of the earlier Skinner vs. Chomsky (1959) debate (e.g., see Galasso 2003, as cited in Owens 2007).1

A central tenet of the debate revolves around morphological storage: viz., how stem + affix process—whether or not stems + affixes are stored as a single item or not (the former constituting a single mechanism model, the latter a dual model). Of course, even Skinner knew that the grammatical construct for regular plural in English was N + s = Pl (e.g., ‘Books’). The question here is not about the spell-out of the grammar, per se, but rather about how such morphological grammars which hold across inflectional morphology get processed. (Skinner’s model on how language is learned would require that e.g., [book] and [books] are rendered as two separate stimuli, as individually coded in the environmental context by ‘singular vs. plural’ respectively: the two items make-up two independent memory schemes, like how go is memorized to change to went).

Proponents for a single mechanism advocate for a full-listing hypothesis which has it that all lexical stems + affixes get processed (stored & retrieved) as a single lexical chunk […]; both regular as well as irregular constructs similarly process as undecomposed chunks insofar that their ‘full-listing’ is directly pulled from out of the lexicon. Such memorized schemes indeed lean towards Skinnerian behaviorism since no decomposed rules are required. Such a full-listing would not only have the items ‘book’ and ‘books’ as two separate (undecomposed) lexical items, but that all derivations of a word would be likewise memorized as separate, independent items—e.g., [speak], [speaks], [speaking], [spoke]*, [spoken] *(noting that a dual pathway would still credit the irregular past tense ‘spoke’ as derived via an undecomposed lexical chunk).

A hybrid model, sometimes referred to as a dual-pathway hypothesis, more commonly known in generative circles as the Dual Mechanism Model, suggest that such undecomposed memorized schemes do prevail with irregulars—where rules can’t be attributed to their construct (e.g., go>went, dream>dreamt, (note ←xvii | xviii→the sound shift), but critically not dream>dreamed which would be a dual model). Many proponents of the dual model even go so far as to compromise (playing nice by debate standards) by crediting a single pathway for what appears by many to be a ‘common regular present tense verb’—for example, the verb ‘do’ seems to fall into the irregular single mechanism route when considering that the verbal present tense inflection of the verb ‘do’ undergoes the same type of sound shifts obliged by a single route of storage & retrieval: e.g., note verb-stem vowel shifts for present tense do>does (similar to past tense dream>dreamt), and, of course, the past tense do>did has all the hallmarks of an irregular verb. Perhaps a phonological clue here to the dual-hybrid-pathway is to say that whenever there is a stem sound shift within the morphological paradigm, a new word must be realized, stored and retrieved in the lexicon as a new single item.

But the hybrid dual-model provides strong evidence that decomposed rules are indeed obligatory for regular stem+affix formations (where no stem sound-shift are observed). I have suggested in various writings that the dual mechanism model can be extended to capturing the dual morphological stages of child language acquisition, where stage-1 memory of undecomposed lexical schemes come on-line prior to a stage-2 of decomposed inflectional rules. Much work of the 1960s generative grammar enterprise, including Berko’s famous Wugs test, set out to demonstrate the nature of such creative and productive morphological processes as only decomposed rules could offer. In this sense, I often use the Skinner v Chomsky debate as a kind of ‘pedagogical device’ in capturing the debate on morphological processing.

←xviii | xix→

This [[…] edge]-feature, as we will show below and throughout the text, allows for the most crucial property of language of all, that of recursion. It is this outside property at the edge which is abstract and removed from the obligatory binding of memory-based schemata. Without such an ability for displacement, movement, affix lowering and raising, inversion, fronting, and most crucially embedding, there could not be a syntax as it manifests in human language.

The Dual Mechanism Model. This dichotomy in essence leads to how a ‘hybrid model’, (i.e., a Dual Mechanism Model (Appendix-2, 5)), which can incorporate both clocks and clouds, might be embedded in our psychological processes of language, and how these two modes of processing indeed have a real physiological presence in our brain: viz., the idea that there are two fundamentally different areas of the brain which bring about this dichotomy of processing. A very simple example of this dual processing could be how an English speaker differently processes the two verbs DO in the expression: ‘How do you do?’, noting how both verbs ‘DO’ have the same spelling and the same phonology, but how they may hold different meanings—of course, also how the two may have very different psychological realities. Try to guess which DO is a clock, and which is a cloud: which has ‘calculative’ meaning and which is ethereal ‘neither here nor there’.

Regarding the Dual Mechanism Model, we can simply note how the expression ‘How do you do?’ allows the deletion of the first ‘do’ but not the second ‘do’ (in quick, spontaneous speech). But why might this be, given that at least on the surface-level, the two verbs appear identical?

Note: Another example of such surface level deception is found in processing distinctions of the Spanish Los vs Las, where the former is an undecomposed memory chunk [Los] and the latter a decomposed stem+affix determiner [[La]s]—with different reaction-time processes: ‘Las’ obtains the signature of rule-based stem+affix dual processing attributed to the movement of affix {s}, while ‘Los’ (being replaced by the masculine determiner ‘el’ within the inflectional paradigm,:*lo/el niῆo, vs la niῆa) reduces to signal as a mere lexical item without movement: viz., where the {s} in [los] rather incorporates into the stem). Spanish speakers don’t process the masculine/plural ‘Los’ as equivalent to the feminine/plural ‘Las’ (despite their surface-level identity). There was a time when ‘lo’ was a functioning determiner in Spanish, but it has since been supplanted by the el>los paradigm: la>las, el>los.

Regarding our ‘clock analogy’, one could say that Spanish speakers no longer count ‘los’ and ‘las’ as being two separate minutes on the same clock. They are on entirely different clocks: one a clock, the other a cloud.←xix | xx→

(Also see processing distinctions between very high-frequency regulars such as ‘walked’ over very low frequency regulars such as ‘stalked’, with high-frequency priming like undecomposed irregular-stem verbs ([go] >[went]). See Ullman, Clahsen).

Well, the two DOs are not identical! In fact, the two have very different psychological states, are located in different areas of the brain, as well as hold different linguistic status. The first ‘Do’ is an auxiliary verb (a cloud) which doesn’t deliver the kind of meaning usually attributed to main verbs, while the second ‘Do’ (a clock) is a main verb which delivers main verb meaning. As we can quickly see, the two verbs hold very different linguistic and psychological status: note the acceptance (in spontaneous speech) of ‘How__ you do?’ versus *How do you__? (where __ marks deletion of ‘do’). One question that naturally follows is if the two different verbs occupy different areas of the brain, in addressing a brain-to-language corollary. They do! The discussion now leads to abstract/non-substantive functional-category words (clouds) such as the Determiner ‘The’, the Auxiliaries ‘Do, Be, Have’ (this is not an exhaustive list), etc. versus substantive/lexical-category words (clocks) such as Nouns, (Adjectives), Verbs (Adverbs), Prepositions, and how the former functional categories are attributed to Broca’s area (front left hemisphere), and how the latter lexical categories are attributed to Wernicke’s area (left temporal lobe).

One question is how this psychological distinction plays out in the trajectory of (i) Child Language Acquisition, as well as in the production of (ii) Second language learning. Particularly, the matter of whether there’s incremental maturational onset of Broca’s area for children—if clouds develop after clocks.2 By extension we can similarly ask if the human Frontal Lobe (FL) of the brain (the ultimate cloud) could have only become neuro-wired later—as a ‘cascading’ consequence—after an earlier neuro-onset of the Temporal Lobe (TL). Linguistics who study ‘Proto-Language’ (Appendix-4) have this progression in mind.

In any case, we know the more robust TL was neurologically connected well before the onset of a FL (both in the development of the child, as well as evolutionarily, and that earlier hominid species may only have had access to TL processes (see Footnote 2)). If so, we might expect functional/Inflectional morphology to go missing in early stages of child language acquisition (2–3years ←xx | xxi→of age). This is what’s behind a maturational hypothesis of child language (e.g., Radford & Galasso 1998).

Priming-effects & Slips. Priming effects are another example of ‘brain as clock’ since the frequency of token items seems to affect how we come to notice and remember certain language structures. We can give a couple of easy examples here:


XXXVI, 278
ISBN (Book)
Publication date
2021 (August)
New York, Bern, Berlin, Bruxelles, Oxford, Wien, 2021. XXXVI, 278 pp., 15 b/w ill., 3 tables.

Biographical notes

Joseph Galasso (Author)

Joseph Galasso (Ph.D., University of Essex) is on the Linguistics Faculty at California State University, Northridge and also lectures as adjunct at California State University, Long Beach. His main research and publications involve issues surrounding early child syntactic development.


Title: Reflections on Syntax