Hello

To me, there is not a single chance to merge the two levels, provided that each one may explain with no exception the other. As you have stated in PPT document, case grammar is but a reflection of the same fact, but from a semantic perspective. How does it differ from that with the syntactic persepctive? The diffremnece is a matter of focus, no more. The noun functioning as subject can be any of agent, recipient, patient, instrument, or theme, etc (of course with varying degrees), and a noun acting as an agent can act in the same situation as patient! In Arabic we say (maata zaydun [ die past masculine singular + Zayd] = literally (Zayd died). In this Zayd is a subject, and supposedly he is the doer of action, i.e., an agent, because he did the action of dying!!!! How? This is a syntactic focus, but in Case Grammar, we say that Zayd is a syntactic subject, but with the theta role of patient. So, considering the diffrent situations in the same angauge, and in diffrent languages reveals that it is useless to try merge the two levels into one.

Regards

Mahdi: You're right. It is just a matter of focus, with one giving priority to syntax and the other giving priority to semantics. I personally would prefer the semantic focus, because it better relates to the real world, but that's only a personal preference. In the end we have to consider Chomsky's concept of elegance: Completeness, Simplicity, lack of ad-hockness, and the generation of all and only the sentences of particular language. Chomsky's generative transformational grammar is now becoming generative Artificial Intelligence: Exciting, but scary.

I don't think that syntax and semantics should be merged. They are complementary and mutually necessary. A syntax without reference to a semantic content would be pointless for natural languages that exist for expressing semantic contents. And viceversa semantic contents need syntactic rules in order to be expressed in a comprehensible way by the speaker and understood in a comprehensible way by the hearer. All this does not mean that the rules of syntax correspond to the rules of semantics.

I agree with Paolo Ramat that syntax and semantics are complementary domains - they cannot be merged, as they operate with different units, but they are connected. In particular, the competition between different syntactic constructions may be often explained by reference to semantic or pragmatic factors.

Paolo and Carlotta: I agree that syntax and semantics are complementary domains. Our brains even tell us so (Broca's area vs. Werneke's area). But there is another issue. Should syntax dominate, or should semantics dominate? And to make this issue even more complicated, what about pragmatics?

Thanks for an interesting conversation. I welcome Don's introducing pragmatics into consideration. A recent book, "A New Semiotics" (Routledge, 2023) by David Sless and Ruth Shrensky dismisses semantics and syntactics as higher ontology, since both are "generated and constructed" through pragmatics (p. 9). As efforts to systematically describe language use, syntax and semantics are important and bring distinct value, but so too does considering pragmatics. For example, context not only significantly helps a speaker or writer to sort out the key issues that most concern people about a subject, propelling a process to persuade ourselves and others about what merits attention, but context is also significant to language choice. Sless and Shrensky describe meaning-making as "a basic form of sharing and creating with others" (p. 21), with meaning created between an "author" and a "reader." Different people give different connotations and denotation to a word, phrase, slogan, etc. It's a convenience to think and act as if meaning is fixed in a word, phrase, image, or action and that everyone uniformly "gets it," much less that they should, or that it is the same "it." Should we be rethinking the relationships among semantics, syntactics, and pragmatics, along with how we talk about communication?

Rodney: Excellent insights. To me it's a matter of higher or lower level of abstraction. Syntax is a lower level of abstraction (more practical); semantics is a middle level of abstraction (also practical, but now more conceptual and less grammatical); and pragmatics (despite the name) is the highest level of abstraction (not grammatical, but very conceptual--even more conceptual than semantics, because more context is taken into consideration). I'm not sure that Pragmatics be in the fieldc of Linguistics. It might be more appropriate in the field of Sociology, Psychology, or even Anthropology). Or even better, maybe more of our research should be cross-disciplinary.

Thank you, Don. We surely could benefit from more cross-disciplinary research. New understandings of communication are likely from collaborations seeking to bridge the divide between scientific and humanistic approaches.

That is a good question. I basically share the idea that syntax and semantics should be construed on an independent basis, this is in line with semantic research. Syntax is not just about encoding the meaning, it is also a system of restrictions (conditions). And semantic differences are not always manifested in syntax, at least overtly. But there is such a thing as semantics-to-syntax interface and some interface models. Fillmore's Case Grammar is a tad old-fashioned theory, and the notions of case role and semantic argument are fairly well assimilated. The principal question is whether the linguists need formal models in general. I think that if a model helps to make some predictions or sheds light on the distribution of some features, it is useful. Taking a case theory more recent and more shallow than Fillmore's, the Marantz-style ternary disctinction of lexical case vs structural case vs defaul case is a dogma, but it is still can be used for explaining some subtle differences in case syntax.

Rodney has written:"syntax and semantics are important and bring distinct value, but so too does considering pragmatics. [...] context is also significant to language choice. . It's a convenience to think and act as if meaning is fixed in a word, phrase, image, or action and that everyone uniformly "gets it," much less that they should, or that it is the same "it."". I fully agree. The semantics-to-syntax interface alluded to by Anton Z. should be extended to pragmatics. The numerous corpora of spoken, spontaneous language(s) we have now at our disposal make it possible to study how language is actually used, what are the strategies speakers make use of in different languages. Therefore, pragmatics is the starting point of linguistic analysis. On the basis of what people do when they speak we can elaborate the morphosyntactic form of their languaging (French 'activité langagière') . Of course, all this is much more difficult when we try to reconstruct the pragmatics of past epochs, known to us only via written texts

Anton: Excellent response. It's OK if the levels remain separated as long as there is eventually an interface between the two. This is the tricky part.

I recommend Don's answer. The same semantic content may be expressed by very different syntactic strategies in different languages. Suffice to think of the multifarious ways of expressing "to have" and of the various ways a language may use for the same content at the pragmatic level.

Case grammar might be seen as a kind of syntactic representation that resembles dependency grammars (or l'actance de Lucien Tesnière) that have proven to be more efficient for machine learning than other "syntactic" representations used for example for translation. But it seems that AI is even more efficient though it is not based on any syntactic or semantic representation. It seems to be based only on statistics.

Script Model Grammar is the best grammar for use in generative AI. Check out the PowerPoint below about "Humor and Computers." I would appreciate receiving any responses, either positive or negative. Thanks.

Paolo: Thanks for your response. Timing is now the issue. I believe that it's now time to start figuring how syntax and semantics should interface. Case grammar is opne possibility. Script grammar is another possibility. The advantage of script grammar is that it can manage grammar loarger than the sentence level, the the paragraph (e.g. the joke) level, the book level, and even the trilogy level. See the PowerPoint about "Computer Humor" in a previous text.