Sorry for the late reply. I have needed some time to think over your scholarly observations.
> (i) I seem to have failed to make clear what I meant by saying:
> >I think we should be surer that Aggava.msa in fact understood it (= tulyaadhikara.ne) in that >way (= having the same substratum).
> Aggava.msa did not coin the term tulyaadhikara.ne, but took it from earlier grammarians of Pali (eg. Kaccaayana) or Sanskrit (eg. Kaatantra). Just from its appearance in 869-871, it isn't clear what Aggava.msa meant by it, regardless of what it originally meant or how later commentaries interpreted it. Perhaps we can draw some conclusions from looking at other places where he uses it.
If you check Kaccaayana (410-412), you will see that the Saddaniiti suttas and their vuttis in this context are but the adaptations of their Kaccaayana counterparts and that Kaccaayana is not clearer either regarding tulyaadhikara.na. This unclearness can be attributed to two alternative reasons:
1. Both Kaccaayana and Saddaniiti did not really understand what tulyaadhikara.na is
2. Or it was not their job to explain what it is.
I think the latter is more plausible (See 4.4 of my uploaded paper).
> I am mindful here of E. G. Kahrs, 'Exploring the Saddaniiti' JPTS 17, pp. 1-212. Kahrs translates Aggava.msa's suttas dealing with kaarakas, and argues, convincingly in my opinion, that although Aggava.msa uses the same terms as in mainline Sanskrit grammar, he does not understand their original sense. ... Something like this could be true of tulyaadhikara.ne as well.
I haven't studied Kahrs' work so I cannot really comment. However, I can at least say that, to compare main Sanskrit grammars with Saddaniiti directly without taking Kaccaayana into account can produce only a lopsided view. For Saddaniiti is only a superset of Kaccaayana; fundamentally they are not really different. (Even the portion you have quoted above is only a variation on Kaccaayana's tune, for instance.)
> I agree that the noun "puriso" refers to a particular man, and that the verb "pacati" refers to a particular action of that man. I am willing to say that that man is the substratum of "puriso", and that that man's action is the substratum of "pacati". But I don't see in what sense that man (alone) is the substratum of "pacati". At best that man is a component of the substratum of "pacati"; how can we say that he is the substratum (by himself)? I much prefer 'referent' to 'substratum' here not only because it is a familiar term, but also because its meaning is clear. The referent of "puriso" is a person and the referent of "pacati" is an action, and there is no way they could be the same. Insofar as it might make sense to say that the substratum of "puriso" is the same as the substratum of "pacati", then 'substratum' and 'referent' must refer to different things.
Your argument is a philosophical one. In this case, at least I can recall the the Neo-logic (navanyaaya) of Indian philosophy, which defines a person's action as his attribute (I cannot give any reference however; all my Skt. books are left at my parents' home in Burma.) However, I think we needn't go that far; the Pali classical grammarians are, as far as I know, not linguistic philosophers like their Sanskrit counterparts. Rather they are only pragmatic linguists, for whom Pali is only a tool, a medium.
Now for the practical side of argument. If we argue that the agent of an action cannot be the substratum of the action, how do we explain even more radical instances, i.e., "ruupe cakkhuvi~n~naa.na.m"? Here the visual form (the object) is denoted as the location of eye-consciousness which is aware of it. In common sense, the object cannot be the location of the consciousness aware of it. However language dictates the linguist, not vice versa; so our job is to try to understand the usage.
In my understanding, the concept of substratum here is nothing but a conceptual tool to explain why there must be concord between the verb and the subject or the object. How? A given action can be viewed from different perspectives. In the case of a man cooking rice, the action of cooking can be viewed a) either as the action of that man b) or as the process of the changes that rice undergoes c) or just as the action itself regardless of the agent or the object. In the first case, the emphasis is on the agent; the active form "pacati" must be used and its agent must agree with it. (usage) Why? because the agent is the substratum of the action (conceptual explanation). In the second case, the emphasis is on the rice; the passive form "paciiyate" must be used and its object must agree with it. (usage) Why? Because the object is the substratum of the action (conceptual explanation). In the last case, the emphasis is on the action itself; so the absolute form "paciiyate" must be used but there is no need to agree with either the subject or the subject. (usage) Why? Because neither the agent nor the object is the substratum of the action (conceptual explanation).
> I can't agree that treating syntax in terms of word-to-word relations is in any sense special to Pali or uniquely Burmese. On the contrary, it seems to me that syntax is nothing more than the word-to-word relations (within sentences). Offhand I cannot think of any approach to syntax of which this could not be said. 'Parsing' sentences in traditional European grammar, or 'tree diagrams' in linguistics are just examples of representing word-to-word relations. Of course that is not to say that different approaches to syntax do not treat or classify such relations differently, or that all grammars put equal emphasis on syntax.
> The matter of word order does not affect this; that some languages such as Pali do not use word order to signal word-to-word relations (while others such as English do) is also nothing special to Pali or uniquely Burmese. There is a large literature concerning the syntactic role of word order in languages; one example is J. F. Staal (one of my teachers). His book 'Word Order in Sanskrit and
> Universal Grammar' (1967) discusses Indian versus Western approaches:
> >Almost all Indian theorists did, either implicitly or explicitly, regard word order as free. For >no independent word is a specific position in the sentence prescribed. Sentences which
> >differ in the arrangement of their words only, are considered as equivalent and synonymous. >(p. 60)
> >For the Indian grammarians, ... grammatical relations between words in the sentence, i. e. >kaaraka relationships and similar grammatical relationships, are expressed by inflexion and >the like. The order of words of the sentence, on the other hand, has no such significance; it >is entirely superficial. (pp. 60-1)
> He is speaking not about Pali here, but Sanskrit.
> Linguists generally regard approaches to syntax as applying not just to one particular language, but to any language. Otherwise we cannot meaningfully compare the syntax of one language with that of another, or understand why children are not programmed to learn the language of their parents, but learn any language they are exposed to while they grow up. From that point of view, Relational Grammar is mistaken in ignoring word order.
I appreciate your detailed and insightful observation. However, I have a couple of reasons to emphasize on the word-to-word relations in contrast to word order in RG.
1. No classic grammar has talked about the word order, and traditionally word order has been treated as a matter of style rather than of syntax. Yet I myself have doubts whether word order is really arbitrary in Pali or in Sanskrit. If it were really arbitrary, a given sentence must be able to appear in all possible variations of word order, and yet still remains the same "sentence". However, when I try to read such variations, I cannot but feel that some variations are better than others even though I cannot explain why. So I personally think RG can be only a pragmatic tool, which cannot guarantee that you can see the full picture of Pali in it.
2. And this is also meant as a warning to Pali learners. In my experience of teaching students of various nationalities, the greatest problem is their tendency to read Pali as they read English; many beginners try to rely upon the word order to get the meaning of Pali sentences. You cannot master RG as long as that mental stumbling block is there.
>I wonder if it has ever been used to analyze Burmese. If so, I >think it would not ignore order.
RG has never been used to analyze Burmese. It is purely a monastic tool to analyze Pali texts.
> In spite of the above, I remain >eager to learn more about the practice of syntactic analysis in >Myanmar, not because of its uniqueness or isolation from other >approaches, but for what it can tell us about Pali. Please continue >your efforts to explain it to us.
Thanks for your appreciation. I will try my best within the constraints of time and circumstances.