1 2
watsup:
Dear members,
I have some confusing questions about syntax and semantics. My friend has introduced me this forum recently and I strongly believe that you all will give me helpful explanations. I would appreciate your help much. These are the problems:
1) what is a grammar in the form of a list of pharse structue rulesneeded to generate all of the below pharse structures rules: a) S-> Det N V Adj N
b) S-> Det N PP V NP PP
c) S-> NP V PP
d) S-> PP V NP
e) S-> NP V NP NP
f) S-> NP V Det N PP
2) what is the difference between grammatical and ungrammatical Englishsentence that CAN be generated by this above grammar, as well as that CAN NOT be generated by this above grammar? Give me examples, please!
3) Are AGENT, PATIENT, RECIPIENT, THEME... the sematic roles of NounPhrases? (depend on verb in the sentence)
4) what is the semantic relationship between colors?

Thank you very much for your all helps. I am looking forward to hearing from you.
This thread originates from within 'usenet', and as such the content and users are not guaranteed to have been moderated by our community.
Don Phillipson:
[nq:1]1) what is a grammar in the form of a list of pharse structue rules needed to generate all of the below pharse structures rules:[/nq]
English has no set of "phrase structure rules"
required to generate one particular set of phrases. English has patterns of acceptable and unacceptable use, and we may be able to derive rules from them
but these remain rules to test phrases as OK or not OK. No rules are "needed to generate" OK
phrases in English.
[nq:1]4) what is the semantic relationship between colors?[/nq]
English has only one semantic rule concerning colours that black is not a colour but the absence of all colours. Semantics has nothing to say about real or imaginary relationships between colours.

Don Phillipson
Carlsbad Springs
(Ottawa, Canada)
This thread originates from within 'usenet', and as such the content and users are not guaranteed to have been moderated by our community.
watsup:
Thank you for your replies.
Re the number 1, what I want to know is that what exactly is "a grammar". Base on what you mentioned, I think, a grammar should be a structure rule showing which components needed to create a sentence and how to form them in the right way. So, I think, the grammar in the form of a list of phrase structure rules needed to generate all of the phrase structure rules listed above should be:
S-> NP (PP) V NP (PP)
NP-> (DET) (A) N (PP)
PP-> P NPSo, a grammatical English sentence can be generated by this grammar might be "A nice girl is eating a cake on the bus". How about a grammatical English sentence that CAN'T be generated by this grammar? Is "Walking down the street is a poor widow" a case? I've learned that "sentences that are impossible because the words are in the wrong order with respect to one another are called ungrammatical" for example "the cat on is the mat" or "the cat on the mat is".

If a sentence is ungrammatical just based on its words' order, so how can be an ungrammatical sentence generated by the above grammar? If I say "Boy jumps", is this an ungrammatical sentence? (the singular noun must have a determiner to be grammatical). If it is, I can create an ungrammatical sentence that can be generated by the grammar above.

Thank you for your explanation in number 4. I know I have go on the right way Emotion: smile
How about the number 4? I am still confused a lot. Can you please clarify me what semantic relationships are? For me, they are still abstractive. Give me examples for what you try to explain to me. I am not a native speaker and actually, explaining something with examples brings me greater opportunities to catch main points. I would appreciate your help very much Emotion: big smile (you see, I know how to use "help" in plural meaning without "s" Emotion: smile, thank you very much)

Please reply me at your earliest convenience. I am looking forward to hearing from you.
This thread originates from within 'usenet', and as such the content and users are not guaranteed to have been moderated by our community.
Don Phillipson:
[nq:1]. . . the grammar in the form of a list of phrase structure rules needed to generate all of ... to be grammatical). If it is, I can create an ungrammatical sentence that can be generated by the grammar above.[/nq]
1. I do not understand the sequence of rules in the wayyou formulated it.
2. You may be confusing grammar (sentence structure) andsemantics (word meaning.) These are normally quite separate and the difference has been discussed at length by many thinkers. E.g. we can see the sentence
The gostak distims the doshes
is probably grammatical (because the verb distim agrees with the siingular noun gostak etc. etc.)
but is void of semantic meaning because the strings gostak etc. are not real words and have no conventional meaning.
The point is that (so far as English really has rules) we can certify correct grammar independently of meaning.
3. You seem to expect the rules (the way you interpretthem) to work backwards as well as forwards. You write
[nq:1]If I say "Boy jumps", is this an ungrammatical sentence? (the singular noun must have a determiner to be grammatical). If it is, I can create an ungrammatical sentence that can be generated by the grammar above.[/nq]
I think you mean that so far as English has rules govern correct sentences the same rules ought to be able to govern incorrect sentences. This is an error in logic. The reality of rules for correct sentences tells us nothing about the reality of rules for incorrect sentences.
This is not a rare or unreal phenomenon. E.g. if we drop a handful of peas onto a plate we may see the peas display certain regular patterns (squares, rectangles, circles etc.) which can be speciified by rules. But other peas fall outside those patterns. However the arrangement of non-pattern peas cannot be described by a set of fules for non-patterns. We have no reason to suppose rules are symmetrical (viz. that for every rule governing XYZ there is another rule governing non-XYZ.)

Another contemporaneous thread asks about the
antonym of a noun chav. But many common nouns
have no (true) antonym, however often they are
semantically associated with pairs. E.g. some people think the antonym of mother is father or the antonym of sister is brother but these definitions generate no reliable rule. (If we derive a rule from mother/father, this rule cannot be applied to sister/brother, and vice versa. This shows the "rule" is defective because not capable of application to other cases.)

Don Phillipson
Carlsbad Springs
(Ottawa, Canada)
This thread originates from within 'usenet', and as such the content and users are not guaranteed to have been moderated by our community.
Alan Jones:
[nq:1]If I say "Boy jumps", is this an ungrammatical sentence? (the singular noun must have a determiner to be grammatical).[/nq]
"Boy jumps" is not ungrammatical. The determiner is often omitted in, for example, the style used for newspaper headlines. "Boy jump" (supposing that "jump" were intended as a verb) would be ungrammatical. This headline style differs in other ways from normal narrative English: in your example, "jumps" may be the historic present, though the full report would be in the past or present perfect tense.
Alan Jones
This thread originates from within 'usenet', and as such the content and users are not guaranteed to have been moderated by our community.
John Flynn:
[nq:1]Thank you for your replies. Re the number 1, what I want to know is that what exactly is "a ... a structure rule showing which components needed to create a sentence and how to form them in the right way.[/nq]
The use of context-free PSGs (or even context-sensitive PSGs) to describe how a natural language works was abandoned a long time ago by most linguists for several reasons, the main ones being that the descriptions started to become very unwieldy and cumbersome, and that the models were simply too linear.
Natural language works with a more hierarchical structure than simply placing phrases alongside each other, which is why most current generative models of language use other means (e.g., binary tree diagrams) to capture the hierarchical dependencies. An element 'higher' up in the structure may have influence on 'lower' elements (or vice versa); this can be seen and explained in a diagram but has no easy way of being shown in a set of PS-rules.
You can get something that looks like a language out of a PSG, but don't expect that grammar to be powerful enough to describe all of a natural language.
[nq:1]So, I think, the grammar in the form of a list of phrase structure rules needed to generate all of ... English sentence can be generated by this grammar might be "A nice girl is eating a cake on the bus".[/nq]
Yes, you've found a set of PS-rules that can produce some output that looks identical to a grammatical English sentence. (Assuming you've defined your lexicon correctly.)
[nq:1]How about a grammatical English sentence that CAN'T be generated by this grammar? Is "Walking down the street is a poor widow" a case?[/nq]
Yes again. You can map the English sentence to your PS-rules but it turns "walking" into a gerund (verbal noun), which is not what the English sentence requires.
[nq:1]I've learned that "sentences that are impossible because the words are in the wrong order with respect to one another ... is ungrammatical just based on its words' order, so how can be an ungrammatical sentence generated by the above grammar?[/nq]
For a few reasons. The main one is semantics. Because it's context- free, it doesn't care what words you use to fill the constituents. It doesn't care if you choose an intransitive, transitive, or ditransitive verb this decision should then select a certain number of obligatory arguments, but the PS-rules don't take this into account, so you can generate weird output: a ditransitive verb with no objects, for instance. Another example, to illustrate another failing, is: - The girl sees himself.
Your PSG could produce that sentence, but I doubt there is any English speaker who would consider that acceptable; the ungrammaticality does not lie in the word order but in which words can be selected by other words/ phrases elsewhere in the sentence.
So, as well as getting all the words in the right order and place, you also have to know what other words are needed by the words you have already chosen or intend to choose. Your PSG doesn't do that, so it's very limited for modelling the full capabilities of English. You could probably fix some faults by introducing context-sensitivity, but that's just a stop-gap until you hit other problems.
You're better off dropping the idea of trying to create a PSG that models the full scope of English grammaticality judgements. Seriously.

johnF
"Ah, but my facts are never factual. (That's a nice quote for John Flynn to haunt me with forever.)"
Robin Bignall, APIHNA, 29 March 2005
This thread originates from within 'usenet', and as such the content and users are not guaranteed to have been moderated by our community.
John Flynn:
[nq:2]. . . the grammar in the form of a ... NP (PP) NP-> (DET) (A) N (PP) PP-> P NP[/nq]
(snip)
[nq:1]1. I do not understand the sequence of rules in the way you formulated it.[/nq]
Um, what's not to understand? The sequence is the traditional "top- down" ordering, starting at S. I admit there are some gaps in it (e.g., a bare V in S instead of building up a VP to give "S => NP VP") but that hardly makes the grammar incomprehensible.

johnF
"What I say in such discussions might not always be easy to grasp , but what I say is never wrong."
Mark Wallace,
This thread originates from within 'usenet', and as such the content and users are not guaranteed to have been moderated by our community.
Don Phillipson:
[nq:1]. . . Um, what's not to understand? The sequence is the traditional "top- down" ordering, starting at S. I ... S instead of building up a VP to give "S => NP VP") but that hardly makes the grammar incomprehensible.[/nq]
Sorry but you may still need to identify what
sort of coding this is.

1. I suppose S = subject and V = verb but Ido not know what NP or PP or DET mean.
Or is V meant to be a disjunction function
(as in logical notation)?

2. It is not clear whether the OP's function ->and JF's function => are the same or different,
and if so how. It is not clear whether either is
an operator or something else.

Don Phillipson
Carlsbad Springs
(Ottawa, Canada)
This thread originates from within 'usenet', and as such the content and users are not guaranteed to have been moderated by our community.
Robert Lieblich:
[nq:2]If I say "Boy jumps", is this an ungrammatical sentence? (the singular noun must have a determiner to be grammatical).[/nq]
[nq:1] "Boy jumps" is not ungrammatical. The determiner is often omitted in, for example, the style used for newspaper headlines. ... example, "jumps" may be the historic present, though the full report would be in the past or present perfect tense.[/nq]
Not to mention that "Boy" is the name of a character in the "Tarzan" books. "Boy jumps" is 100 percent grammatical if "Boy" represents that character.

Bob Lieblich
You Jane
This thread originates from within 'usenet', and as such the content and users are not guaranteed to have been moderated by our community.
Show more
Live chat
Registered users can join here