The Origins of Junction Grammar

Site  sponsored by Linguistic Technologies, Inc.


Eldon Lytle
Linguistic Technologies, Inc.


A flashback will perhaps be helpful in putting Junction Grammar (JG) in the context of its origins.

Transformational Grammar

The 1960’s were, to say the least, an era of national hand-wringing and personal conflict in the United States. Not only was the war in South Vietnam running its deadly course but the war in linguistics was heating up. The latter conflagration, while perhaps not making headlines, was pernicious and deadly in its own right. Linguistic science on the American scene had been assaulted and overrun by the warriors of the so-called Chomskian revolution, a subtle application of behaviorist psychology to linguistics that bad-mouthed Skinner’s take on ‘verbal behavior’ while, for all practical purposes, retaining Watson’s aversion to introspective analysis. Thus, in putting forth Transformational Generative Grammar (TGG), Chomsky argued (correctly) for the proposition that language is largely innate (rather than conditioned) but declared ‘appeals to meaning’ (semantics) as being unnecessary to characterize that innateness. After all, Chomsky mused, what could hair color possibly have to do with constructing a grammar? [1]

Semantics as an Afterthought

Intrinsic to the entire approach, of course, was a reluctance to deal with semantics, a turn of mind fostered by the behaviorist notion that studies of meaning were subjectively tainted while studies of syntax were not. The result was a model of ‘syntax’ that deployed a set of rewrite rules (P-rules) to generate ‘deep structures,’ and another set of rules (T-rules) to transform those into the ‘surface structures’ that we observe overtly. Semantics, for its part, inasmuch as it refused to go away, was subsequently appended to the model as an afterthought in the form of an interpretive function. Let’s hear it from Chomsky himself: 

    The generative grammar of a language should, ideally, contain a central syntactic component and two interpretive components, a phonological component and a semantic component. The syntactic component generates strings of minimal syntactically functioning elements (following Bolinger, 1948, let us call them formatives) and specifies the categories, functions and structural interrelations of the formatives and systems of formatives. The phonological component converts a string of formatives of specified syntactic structure into a phonetic representation. The semantic component, correspondingly, assigns a semantic interpretation to an abstract structure generated by the syntactic component. Thus each of the two interpretive components maps a syntactically generated structure onto a “concrete” interpretation, in one case phonetic and in the other, semantic.[2]

Paradigm Politics

As a grad student working towards a Ph.D. in Slavic Languages and Literature at the University of Illinois (Champaign-Urbana), my first exposure to the ‘new linguistics’ occurred in the fall of 1966. Several of Chomsky’s protégés were practicing their craft in the linguistics department there at that time. They were supremely confident in the proposition that no plausible alternatives to Transformational Grammar could possibly exist. Not content to peddle their own wares, some set about to publicly attack and discredit the work of any and all not subscribing to Chomsky’s gospel.

A particular case comes to mind: A well-respected field linguist, having devoted some twenty years to the task of reducing a preliterate language to print, delivered a report of the work in a heavily attended presentation sponsored by the Linguistics Department.  An emissary from the TGG consortium strolled in towards the end, only to preoccupy himself with a book. He punctuated his reading, however, with intermittent audible remarks to persons near by as the report wound down. Responding to the invitation to ask questions, this person was the first to have one. “How does it feel to have wasted twenty years of your life on something so utterly uninteresting and trivial?” Having delivered the insult, the questioner then walked out. 

If one is to believe reports circulated at the time, reprehensible behavior of this ilk repeated itself elsewhere. In retrospect, it appears that a deliberate, orchestrated attempt was made to discredit persons not subscribing to TGG by attacking and insulting them in public forums. As the movement gathered momentum, it took control of major journals and organizations, including the Linguistics Society of America. Thereafter, submissions for publication in its journal not in accord with TGG dogma were returned to their authors with the dismissive pronouncement that they were “not reflective of mainstream linguistics in America and therefore not of interest to our readers.”  The ultimate objective, it seemed, was to co-opt funding, which, to the detriment of linguistics at large, was largely achieved. The time came, in fact, when professors of linguistics not affiliated with Chomsky’s juggernaut ran the risk of getting a pink slip. It was a liability to have them on staff.

‘I Have a Question’

Meanwhile, in the classroom, which is where I was, it was risky business to question any aspect of THE MODEL OF SYNTAX.[3] “Why,” I ventured to ask, “since deaf mutes and even some animals evince linguistic behavior, is ‘deep structure’ constituted with verbal formatives?” No satisfactory answer was forthcoming. “Given that one’s primary intent in speaking is to ‘mean’ something, wouldn’t it make more sense,” I asked, “to make semantics generative and syntax interpretive rather than the other way around?” With that, it seemed, I had crossed the line into never-never land. “You’re obviously not intelligent enough for this course. I’m recommending to your committee that you change your major.” Subsequently, my term paper detailing the manner in which sentence structure (syntax) was coupled with voice intonation was summarily rejected. The party line, I discovered, stipulated that syntax and phonology were ‘autonomous.’[4] 

Needless to say, students daring to do their own thinking found one barrier after another thrown up to impede progress towards completion of their Ph.D. programs. The prime instrument of control was the dissertation.  Either it was ‘conformative’ … or no degree.  It took a letter documenting the obstructionist tactics to the department chair accompanied by a transcript of grades (essentially straight A’s) - all copied to folks in D. C. in charge of the National Defense Title-IV scholarship program - to break things loose for me.  After that it was clear sailing. I straightway constructed a model of language that jived with my own intuition and was actually allowed to use it as the basis for that all-important dissertation.[5]


Meanwhile, with battle lines drawn and transformationalists firmly in control of professional linguistic institutions, the opposition began to re-institutionalize.  The Linguistic Society of America under Chomsky’s influence had not only squelched alternative views but actively supported the anti-war movement of the 60’s. The exodus of LSA members in the wake of these indiscretions eventually gave rise to the Linguistic Society of Canada and the United States (LACUS), which, under the watch care of Adam Makkai and his colleagues, steered clear of political entanglements and provided a friendly, open forum for the free flow of ideas and discussion.[6]


Centering on Semantics

Such were the origins of Junction Grammar (JG). How is it different and why is it important? JG was a gut reaction to the manner in which the roles of syntax and semantics were defined in TGG.  In sidelining semantics, Chomsky had set up the syntactic component as the generative base and assigned to semantics an interpretive role. JG turned them the other way, making semantics generative and syntax interpretive.[7] 

The semantically-oriented generation rules (J-rules) of JG were drawn up as algebraic expressions having both operators (junctors) and operands (sememes). Junctors were defined as falling under three major operational categories, namely, (1) adjunction, (2) conjunction, and (3) subjunction. This formulation immediately bore fruit, yielding in schematic form what functions, in essence, as a Periodic Chart of Constituent Structures (PCCS). To make a long story short, the formalism of the JG base quickly morphed into the discovery procedure that had for so long eluded linguistic science. Those were days never to be forgotten, for, as Einstein so aptly expressed it:

    It is a wonderful feeling to recognize the unity of complex phenomena that to direct observation appear to be quite separate things.[8]

Lexical Coding

Lexical-coding grammars consisting of L-rules (lexicalization rules) transposed (encoded) the semantic structures of the JG base into corresponding lexical strings.  Lexical phenomena (including word order) were thus abstracted away from the base and repositioned under language-specific L-rule components.  Under this arrangement, outward structuring (syntax) is a coded reflection of universal structuring with a particular language-specific overlay. The meld of universal and specific elements was thereby accounted for while simultaneously providing an account of their respective origins. 

The transformationalist, therefore, in reading this collection, will kindly bear in mind that JG is a coding model – not a derivational model.  In JG, language-specific coding grammars interpret ‘deep structures,’ which are its semantic representations, in the process of constructing verbal symbolism to represent them.

Biological Foundations

Gerald Edelman, director of the Neurosciences Institute and chairman of the Department of Neurology at the Scripps Research Institute, has roundly criticized TGG for making only marginal reference to the biological foundations of the mechanisms it purports to explain. Its proponents, he asserts, envision mental software that does not exist a priori and then claim that it doesn't matter what hardware (neuralware?) it runs on.[9]  From his perspective, modeling language in the abstract is a deviation of science for which there is no excuse. 

Having been conceived with biology in mind, Junction Grammar shares Edelman’s aversion to ‘mythological linguistics.’  We quote from a lead article in this collection: 

    Linguistic data types appear to be similarly constrained by virtue of the functional diversity of the tissues and organs which constitute the human capacity for speech. For example, code designed to activate the musculature of the vocal tract could neither be used to drive the writing hand nor to stimulate the neurological tissues of any "semantic" tract. From the point of view of junction theory, then, it would be unnatural to fuse these codes as though they belonged to a single representational system. Rather, it would be necessary to maintain them separately, in order to satisfy the unique content and formalism of each, and then, additionally, to provide a means for them to interact if necessary. 

    The failure of linguistic science to do this systematically is largely responsible, in my opinion, for the embarrassing absence, at this late date, of the at least relatively unified base which other more prestigious areas of scientific inquiry enjoy. Some will counter, perhaps, that linguistics is a young science, and that the present diversity is healthy. 

    Excuses aside, I am firmly convinced that the crux of the problem is rooted in the naive practice of "cross-breeding" data types in unnatural ways. While the resulting offspring are plentiful, and, to be sure, captivating in their infancy, and promising in their youth, they have a way of aborting prematurely, or developing fatal post-natal deformities. Still others mature as impotent mules, as it were, or as "linguicorns" which have no counterparts in reality, but exist only in the world of what might be termed "mythological linguistics."[10] 

Correspondingly, the JG model of language deploys components intended to reflect specific anatomical assets that language necessarily engages. The interfacing between modules is strictly a matter of transposition (as opposed to derivation). Neurobiologists are invited to examine JG’s modules with reference to their own research and comment on correspondences or the lack thereof with their own findings.

In a Nutshell

In sum, Junction Grammar does not transform its ‘deep structures.’ JG’s base rules (J-rules) are not a passive vehicle for characterizing linguistic ‘competence’ in the abstract. They are, rather, intended to model the ongoing process of information management that transpires inside our heads. In a netshell

  • Each J-rule, as it executes, accomplishes what no other rule could.  In their dynamic, proactive role, J-rules are seen as responding to specific needs that are met by precisely the composition of rules brought to bear at a given point in time.  To transform them in any way would thwart their intended purpose.


Mixing is Bad

If verbal formatives are present in the base, as they are in TGG, they propagate outward with troublesome consequences. From the JG perspective, as noted in the foregoing paragraphs, many of the cruces of TGG are seen as arising from the attempt to deal with what amounts to an amalgamation of semantic and lexical phenomena in a single representation, which, in the vernacular of computer science, mounts up to a flagrant data-type violation error. The rational mishmash thus created has given birth to a rather motley family of amended models, including ( but not limited to) Structural Grammar, Relational Grammar, Word Grammar, Autolexical Grammar, Construction Grammar, Functional Grammar, Categorial Grammar, Head-Driven Phrase Structure Grammar, Lexical Functional Grammar, Systemic Functional Grammar, and so on.[11] Each is responsive to some problem, but none to date, so far as I know, has gotten it right. For that, one simply has to abandon mainstream architecture and build up a new edifice with corner stones in proper orientation with respect to one another.


That, of course, is what JG was, and is, all about. The present scene in linguistic science is nothing less than scandalous! At this preposterously late date, its practitioners yet labor in what amounts to a ‘warm-up’ state, akin to ‘an orchestra expectantly tuning up, each player absorbed in his own instrument only, deaf to the caterwaulings of the others.’[12]  Such need no longer be the case. JG can do for linguistics what the Periodic Chart of Elements did for chemistry and Schrödinger’s equation did for physics. It’s incisive structural generalizations lead not only to representational elegance in the laboratory but to practical, intuitive applications in the classroom.

It all comes down to this: I am convinced that when the valid elements of the currently competing models finally come into focus, they will be found to coincide with the essentials of Junction Grammar. These essentials are:

    I. Architecture

    • A network of specialized, biologically-grounded components interconnected by means of coding interfaces.

    II. Notation

    • Natural categories (labels) deployed in each operational domain as operands.
    • For each component, an array of medium-specific joining operations comprised of three basic types, namely, conjunction, adjunction, and subjunction.
    • An algebra for each component, the expressions of which consist of the categories and joining operations (junctions).
    • Context-sensitive procedural statements that transpose (encode and decode) between components.

For detail, the reader is invited to read the articles included here. For advanced detail, the reader is directed their sequel, LANGUAGE in Capital Letters.

Eldon Lytle
Rose Valley, Nevada
December 12, 2003 


[1] Noam Chomsky, Syntactic Structures (The Hague: & Company, 1957), p. 93.

 [2] Noam Chomsky, Current Issues in Linguistic Theory (The Hague: Mouton & Company, 1966), p. 9.

 [3] The sequel to Chomsky’s initial exposition of his theory was entitled ASPECTS OF THE THEORY OF SYNTAX (Cambridge, Mass: THE M.I.T. PRESS, 1965).

 [4] To his credit, this particular professor later approached me at a conference at Indiana and apologized for his behavior. In the interim, it seemed, Chomsky had retooled his position on the relative ‘autonomy’ of components in his model.

 [5] Eldon G. Lytle, Structural Derivation in Russian. Ph.D. Dissertation, University of Illinois, Champaign-Urbana, 1972.

 [6] Eldon G. Lytle, “JUNCTION GRAMMAR: THEORY AND APPLICATION.” Invited lecture presented at the annual forum of the Linguistics Association of the United States and Canada (LACUS), August, 1979. Included here in this collection. 

 [7] In adopting this approach, JG ‘deep structures’ joined company with what latter came to be called ‘language of thought,’ or ‘mentalese.’

 [8] Extracted from a letter to Marcel Grossmann from Albert Einstein. As quoted in Wilson, Consilience: The Unity of Knowledge.

 [9] Gerald M. Edelman, Bright Air, Brilliant Fire: On the Matter of the Mind. (New York: Basic Books, 1992.)


 [11] The term grammar is often used en lieu of model in linguistic literature.

 [12] Arthur Koestler, TheSleepwalkers: A History of Man’s Changing Vision of the Universe (London: ARKANA Penguin Books, 1989 ), p. 25. Koestler uses this analogy to depict the state of cosmology before the advent of Pythagoras as the concert master of the ancient Greeks.

 [13] Eldon G. Lytle, LANGUAGE in Capital Letters (Las Vegas: LingTech, 2003). Ebook version online here.

[Home] [Origins] [Article Archive] [Forum] [Guest Book] [BYU]

Copyright© 2004 Linguistic Technologies, Inc.