This document has a long and tragic history, recounted below for any who might have an interest in it. Meanwhile, it is 

Available in PDF format: (5M+)

(Note: This is a pre-final version, subject to change as errors are noted or reasonable enhancements come to mind.)

History

This dissertation was defended in December 1981, at the University of California, San Diego. It contains one of the earliest presentations with any pretensions to comprehensiveness —perhaps the earliest by someone other than Ronald Langacker— of the Cognitive grammar model. It uses that model for the analysis of a sizeable chunk of what is perhaps the most central and complex part of any Aztecan language, namely the verbal morphology. The particular Aztecan variant expounded on is Tetelcingo Nahuatl (ISO code nhg), also known as Mösiehuali, spoken in Tetelcingo, Morelos, Mexico. 

The original title spoke of an exploration in Space grammar rather than Cognitive grammar; it was not until several years later that the model began to be called Cognitive grammar. 

It was written on the UCSD computer, formatted for the troff program as I remember, typically in sessions from 2 or 3 am till things got crowded in the computer lab about 9 am. It was printed out on a daisy-wheel printer with special daisy-wheels that had a few more of the characters I wanted than the standard ones did. I wore several of them out, printing preliminary versions. Two special aspects of the printing process were the need to program in fore-and-aft stops wherever (happily in few places) Greek characters appeared, in order to change daisy wheels; and the shenanigans I put the poor printer through to produce my special characters and other special effects. For instance, to print a ƛ, I made it print an underlined \, backspace and print a - (though that was omitted in the final version), backspace again, scroll down a certain amount and print a ^, then scroll up and continue whirling on its merry way. At daisy-wheel speeds, I probably added an hour or more to the print-out time by that little trick. The diagrams were laboriously drawn by hand with felt-tip pens and plastic templates, and again laboriously (and most literally) cut-and-pasted into the pre-final version which, after certain phonetic letters and diacritics were written in by hand, was photocopied onto the final version.

When I left UCSD I took with me a mighty mag tape some 20 inches in diameter, containing not just my dissertation files but all the computer files for papers and computer programs that I had generated during my stay there. Within three years (or was it four?), the computer center at UCSD itself could no longer read tapes in that format. I shudder to think of what intellectual treasures were thereby lost to posterity, and irreplaceably momentous software alas. (I think I finally pitched the tape in the mid-90’s, having given up on ever finding a place that could copy the data onto some kind of modern media like hardshell floppies.)

SIL, with whom I had been working since 1970 (and still do work) at some point inquired about publishing the dissertation. I was amenable to the idea, but realized it would have to be recreated from the printed copies. In the early-to-mid-90’s, Ramona Millar (God rests her soul) typed it for me, on a pre-Windows DOS system, as I remember, with Standard Format Markers flagging the more obvious formatting changes. At various times during the years since then, I have worked on proofreading, formatting, constructing tables, redrawing diagrams (a daunting task), etc., to come up with the version here presented to you. 

Differences from the original version

There are several sorts of differences from the original version, including enhancements made possible by more modern technology and (relatively few) editorial changes —all improvements, it is to be hoped.

The editors of the SIL dissertation series (who may yet publish the thing!) sent it out for a reviewer to suggest changes. The major change that was requested was to change the name of the linguistic model from Space grammar to Cognitive grammar. Since Langacker had been prevailed upon to make that name change in the mid-1980’s, and the old name has largely disappeared from the consciousness even of those who once knew it, this seemed a reasonable request. 

It was gratifyingly (or perhaps, depending on your theoretical predilections, deplorably) clear that the model had changed very little from the time I wrote the dissertation to the publication of the (then-)definitive presentation, Langacker’s two-volume Foundations of Cognitive Grammar (1987, 1991). Indeed, the vast majority of what I had written in 1981 still seems to me to be correct and not superseded (though often enriched or legitimately viewed from other angles) in subsequent work; it is quite thoroughly compatible with even the latest presentations of the model (e.g. Taylor 2002, Langacker 2007.) In any case, I did not feel I could easily update the analysis without embarking on a complete re-write of the whole thing, turning it into a rather different document. It seemed better to leave things as they were except for minor adjustments. They are of the following kinds:

• The name change from Space grammar to Cognitive grammar (and the abbreviation SG to CG).
• Correction of typos and other clear errors and inconsistencies.
• Rewording of (some of the) places where I had gotten tangled in my own syntax —easy for me to do— or had otherwise managed to make myself unclear. A number (by no means all) of my absurdly long and convoluted paragraphs and sentences have been broken up as part of this process.†
• Updating of references to subsequently published works.
• A few comments, included between square brackets, regarding matters where what I wrote now seems to me to be wrong or misleading.
  
• Redrawing of the diagrams on computer —I used various versions of Visio, starting in its pre-Microsoft days—, with the use of color to mark profiles, active zones, and the Trajector/Landmark distinction. (This was thoroughly orthodox, a natural technological development from Langacker’s celebrated “Space Grammar chalk”. I had omitted it in the original only because color copies were impossible for me.)
• Abandonment or adjustment of the originally mandated stylistic dicta regarding  margins, line-spacing, the formatting of section headings, and so forth.
  
• Use of proportional fonts, tables,  “pretty quotes”, and other such enhancements.
• Use of different fonts or font enhancements for marking different kinds of material. (In the original I had used underlining, and inclusion in a box, for similar purposes.)
• Use of Unicode characters, which permitted reference to IPA characters in phonetically related discussions, etc.

It is my hope that this document may be of use to those interested in Cognitive Grammar, including the history of the model; to those interested in Aztecan and in Tetelcingo Nahuatl in particular; and to those interested in the many types of lexical, morphological and syntactic structures described in its pages.

Langacker, Ronald W. 1987. Foundations of Cognitive grammar: Vol. I, Theoretical prerequisites. Stanford: Stanford University Press.
——. 1991. Foundations of Cognitive grammar: Vol. II, Descriptive application. Stanford: Stanford University Press.
——. 2007. “Cognitive grammar.” In Geeraerts, Dirk, and Hubert Cuyckens, eds., The Oxford Handbook of Cognitive Linguistics, p. 421-462. Oxford: Oxford University Press.
Taylor, John R. 2002. Cognitive grammar. Oxford: Oxford University Press.

Note:
† Oh, well, at least I am relatively safe from Alzheimer’s.