The concept of Grammar

Giorgio Buccellati – February 2016, January 2025

Encoding manuals as envelopes

The antecedents of the Grammar are to be found in the various Encoding Manuals that governed the recording system since the beginning of the Urkesh/Mozan Archaeological Project. They are of interest in that they shed light on the development of the system. But they are also interesting in order to highlight the nature of the Grammar.

By its own nature, an encoding manual has two aims:

to provide a list of definitions that are available for describing the items known to belong to a certain inventory, and
to resolve the codes that are used to formulate these definitions.

In an encoding manual, the focus is on the details. There is, to be sure, an overall cohesiveness that governs the list of definitions and the corresponding codes. But this is in the nature of providing an envelope, or a set of envelopes, that neatly assemble these discrete elements which would otherwise be floating in a vacuum.

A trend beyond

It may be said that an encoding manual tends to be a grammar, since, for the “envelope” to be effective, it must clearly delimit what it may, and may not, contain. This implies, however subtly, a notion of the whole, but it is truly only an implicitation, i. e., it is not a declared principle: an encoding manual does not articulate the conceptual and theoretical framework that defines the system as a whole.

It may also be said that an encoding manual has a hidden theoretical component. In the case ot the Mozan/Urkesh project, this was briefly enunciated, in a set of “Principles and presuppositions,” of which one will find in the UGR digital book the 1992 version.

Even more important in the development of the Mozan/Urkesh encoding manual, there was the presence of overriding concept of a system: this served as the underlying buttress that kept the system coherent over the several decades of operation though which it evolved. It is for this reason that data entered in the early stages of the project are still usable within the current framework.

Grammars as systems

It is clear, then, that the Grammar is not an overgrown encoding manual, but something substantially different. The main difference is that a grammar emphasizes the structural coherence of the whole. It does certainly include the details that are at the core of an encoding manual. But these details are seen in function of the whole.

This may best be explained in terms of the difference between a list and a paradigm. A list is not a closed system: it is open to any addition that may fit at any place in the sequence. A paradigm is instead a closed system: any change or addition entails a revision of the system as such, a revision that affects by necessity all the specific cases where the paradigm aplies. If the paradigm of a verbal tense admits three persons win the singular and three in the plujral, adding a dual voice implies that all verbs must accept it. There can be no ad hoc and casual addition that pertains to only one or some lexical examples.

An example

Aa sn example I may refer to the paradigm of contact associations. The “list” given in the main lexicon is indeed a closed system: it proposes ten possible “verbs” which form a tight paradigm, aiming to include all possible types of contacts, organizing them in a specific and tightly knit arrangement of five categories, each subdivided into two alternatives. The actual occurrences in the excavations are given in the sitewide digital book.

The paradigm does not exclude additions: but any such addition would have to show how it affects the other elements in the system: is it a refinement of an existsing category? a totally new category, and if how does it relate structurally to the others? and so on.

Database encased in an argument

In a database, however relational it may be, we have a repository of facts (the “data”): they are, indeed, coded so as to show not only what the deifinition of each item, but also its potential connection with others. But the connection remains, precisely, potential. An ecoding manual serves well such a purpose.

The grammar goes beyond that: it is thus intended to serve the development of an argument, and not just to serve as the presupposition of a database. An argument aims to see the data as part of a structural whole, where the data are seen in relationship to each other and in function of what this relationship means. They must lead from a premise to a conclusion, according to a sequential logic. And a grammar, being constructred in function of such a whole, can

The concern for seeing the coding system as a grammatical structural whole is what makes possible on the one hand the development of the segmented narrative, while on the other it allows for the full interaction among all levels or planes within any given website and among websites.

Linguistic model

Power

The quality of a grammar is judged in part by its “power,” i.e., the effectiveness with which the rules generate an output.

The actual process whereby the output presented here is generated from the data represents an altogether different exercise, requiring as it does the application of specific input protocols and the running of a set of programs. The development of these protocols and the writing of the programs has occupied much of my effort, alongside the development of the theoretical framework embodied in the Grammar: they are described in detail in a separate website, the Digital Operation Manual

While actual use of protocols and programs obviously requires a certain amount of training, two main points may be stressed here:

The protocols are intuitive, and the programs have been written in function of their use. As a result, the actual recording process, and the coterminous process of data entry, are typically learned quickly and have typically become second nature for the staff.
The programs operate with great speed on any normal computer, so that the buildup of the entire record for any given excavation "book" takes place in contant concomitance with the data entry itself. Typically, it takes a few seconds to process the few files that are produced on any given day, so that at the end of the day we have in effect the final publication.