Theiling Online    Sitemap    Conlang Mailing List HQ   

Stack-based syntax (was: affixes)

From:Ray Brown <ray.brown@...>
Date:Sunday, February 20, 2005, 19:09
On Saturday, February 19, 2005, at 11:01 , # 1 wrote:

> Mark J. Reed wrote: > >> I don't know Fith, so I'll create an example with English words and just >> put them together postfix style. Prefix notation is sometimes called >> "Polish notation" because a Polish mathematican first proposed it, and >> postfix notation is therefore called "Reverse Polish Notation", or "RPN" >> for short. So let's call this syntax variation "RPNglish." > > OK I'll try to reformulate your explaination just to know if I understood > > In your mathematical example with mathematic operators, each of them where > to link 2 of them but in a language they operate on different number of > operable arguments (nouns/pronouns)
They can, but do not have to. IMO if one is going to produce a language à la Fith with a stack-based syntax, this is done most efficiently with binary operators.
> there are some that go only on one to give a precision about the meaning
> Those are probably all the kinds of adjectives: descriptive, numeral, > possessive, demonstrative, indefinite. They may describe the noun, modify > it, define it, or tell other precisions such to whom it belongs
Three points here: - the category 'adjective' is by no means universal. As Trask says of adjectives "A lexical category, or a lexical item belonging to this category, found in many, though not all, languages...." - you are lumping together into one category groups of words which are trated differently in some languages. - in any case, I would consider many of these items (certainly descriptive adjectives) to be bound to their noun by a binary operator, say: beauty, city, REL = a city [which is] beautiful/ a beautiful city. I am strongly of the opinion that if one is going to construct a truly stack-based syntax one simply has to think in terms quite different from traditional western 'parts of speech'.
> Articles would also work that way to tell the definiteness of the noun
_Many_ languages do not have articles. I believe in fact the majority do not. So I think one needs to ask whether they would figure at all.
> Case marks would act that way with a single noun
Surely not! Is not the whole point of case markings that they show how the noun/pronoun _relates_ to some other part of the sentence (usually the verb). As I see it, the the case markings are binary operators.
> And the intransitive verbs also because they only affect one noun
I do not see how a lexical category like verb should act as an operator. Isn't it something more like: singing, John, NOM = John is singing?
> Other operators will simply link two arguments in different ways like > conjonctions and other in more sophisticated ways like transitive verbs.
I would see a transitive verb more like: (John NOM (loving ACC Jennifer)) which in postifix (stack-based) form would be: John, loving, Jennifer, ACC, NOM. [snip]
> Some operators link 3 arguments like ditransitive verbs
IMO ternary operators are neither desirable nor necessary (probably spent too many years teaching computer science :) (John NOM( giving ACC ( book DAT Jennifer))) --> John, giving, book, Jennifer DAT, ACC, NOM.
> On verbal operators, I don't know how would be indicated the TAM,
Yes, that does add complications :) But it depends rather what you want to mark and why. Many languages get by, for example, without explicit tense marking.
> would they > be forced to be affixed on the verb or should there have a third level of > operators to affect the other operators?
I am not sure what you mean by 'operators to affect other operators'. How would that work? =============================================== On Saturday, February 19, 2005, at 11:56 , Mark J. Reed wrote:
> On Sat, Feb 19, 2005 at 06:01:53PM -0500, # 1 wrote:
>> Case marks would act that way with a single noun > > As with other syntactic models, you can mix analysis and synthesis. A > case mark could either be part of the noun, or be a unary operator > ("unary" = "takes one argument").
Surely a case marking, whether by suffix or by a post-position (in a stack-based syntax it must one or t'other), is a _binary_ operator as I explained above?
> As an example of this in the > mathematical world, a negative number like "-123" can be considered as > an argument (the - is just part of the number), or as the unary > operator "-" (negation) applied to the argument "123", which would be in > postfix something like 123-. But since using "-" for both subtraction > and negation would be ambiguous, most postfix systems either allow forms > like "-123" and treat a - followed by a digit differently from a - by > itself, or else use a different symbol - for instance, negative numbers > in the UNIX program dc are marked with a _ instead of a -.
Yes, indeed. The unary NEG must be treated either differently from the binary SUB. Even if one allows -123 to be treated as a single literal distinct from 123, it is difficult to see how all unary operators can be avoided; for example logical NOT must surely be treated as a unary operator. [snop]
> Again, it could be either analytic (affixed) or synthetic (an operator).
Puzzled by this. Surely, whether one a a synthetic suffix or an analytic postposition, the thing is still an operator in a stack-based system? ==========================================
> On Sunday, February 20, 2005, at 03:56 , # 1 wrote:
> Isn't exactly as an SOV, postpositionnal, noun-adj language?
IMO no - not exactly. Certainly such a language might be a good place to start. [snip]
> As I see it, a LIFO grammar is only a complicated way to explain a grammar > that is explainable in traditional way?
No. A language with a truly stack-based syntax would not IMO be easily explained in terms of traditional western grammar.
> > A sentence like: > > dog the big cat your small love = the big dog loves your small cat
But this doesn't clearly separate lexical items and operators. This would imply, for example, that 'love' is a combination of lexical item and operator. IMO in a truly stack-based system, lexical items and operators should be kept distincr.
> could be an SOV, postpositionnal, noun-adj language > > The article after "dog" means it is definite,
Yes, one could accept THE as a unary operator, but...
> the possessive after "cat" > means who owns it,
Quite - *who owns* - two concepts here. 'your' will do as a single item. It is surely "cat you GEN"?
> the adjectives are placed after the nouns and the verb is > at the end of the sentence > > But it could also be a LIFO sentence with "the", "your", "big", and > "small" > being simple operators
IMO neither 'big' nor 'small' can be _simple_ operators. They are surely lexical items and need an operator to link them to some other lexical item.
> and "love" being an operator to link the two > preceding arguments
> > Is there a difference I didn't see??
I think so. ==============================================
> On Sunday, February 20, 2005, at 06:07 , Tristan McLeay wrote: > > On 20 Feb 2005, at 8.01 am, Mark J. Reed wrote:
>> Anyone past middle school has learned that there is a standard order of >> operations in such expressions. In English, we learn the phrase "My >> Dear Aunt Sally" to remember that Multiplication and Division come >> before Addition and Subtraction.
> I've never heard 'My Dear Aunt Sally' before; well, not in reference to > maths.
Nor I.
> I was taught in grade five the more complete 'bodmas'; my sister > was taught 'Bomdas' (but 'Bodmas' seems to have been more common; my > subsequent two schools,
Bodmas is what I was taught here in the UK - and my computing students still talked of Bodmas. I guess 'My Dear Aunt Sally' is an Americanism : ) [snip]
> It stands for Brackets, Order, Divide, Multiply, Add, Subtract;
Precisely - and that's the weakness of infix notation, without some conventional rule of 'order of precedence', the expressions are ambiguous. Whereas in both prefix and postfix notations, expressions are unambiguous and we need neither Aunt Sally nor Bodmas :) =========================================== I too know very little about Fith, except that the name was inspired by the programming language called FORTH which uses 'Reversed Polish' or post-fix notation. If I was challenged to produce a language stack-based syntax, I would certainly use a model that confined itsel to binary operators with, maybe, a few unary operators. It occurs to me that Lin's "cements" are indeed binary operators. It even has order of precedence in that 'internal cements' bind more closely than 'external cements'. Indeed Lin could easily be transformed into a stack-based system, e.g. u_f+h --> ufh+_ (You see the bird). IIRC the operators in Tom Breton's AllNoun are binary. Um - I had better look out my notes. IMO with a stack-based syntax, lexical items are literals and operators make explicit how the literals relate to one another. Ray =============================================== =============================================== Anything is possible in the fabulous Celtic twilight, which is not so much a twilight of the gods as of the reason." [JRRT, "English and Welsh" ]