mirror of
https://github.com/GrammaticalFramework/gf-core.git
synced 2026-04-10 13:29:32 -06:00
1197 lines
37 KiB
HTML
1197 lines
37 KiB
HTML
<!DOCTYPE HTML PUBLIC "-//W3C//DTD HTML 4.0 Transitional//EN">
|
|
<HTML>
|
|
<HEAD>
|
|
<META NAME="generator" CONTENT="http://txt2tags.sf.net">
|
|
<TITLE> Constructors: the Resource Syntax API</TITLE>
|
|
</HEAD><BODY BGCOLOR="white" TEXT="black">
|
|
<P ALIGN="center"><CENTER><H1> Constructors: the Resource Syntax API</H1>
|
|
<FONT SIZE="4">
|
|
<I>Last update: 2007-07-08 16:44:35 CEST</I><BR>
|
|
</FONT></CENTER>
|
|
|
|
<P></P>
|
|
<HR NOSHADE SIZE=1>
|
|
<P></P>
|
|
<UL>
|
|
<LI><A HREF="#toc1">Principles of organization</A>
|
|
<LI><A HREF="#toc2">Texts, phrases, and utterances</A>
|
|
<UL>
|
|
<LI><A HREF="#toc3">Text: texts</A>
|
|
<LI><A HREF="#toc4">Punct: punctuation marks</A>
|
|
<LI><A HREF="#toc5">Phr: phrases in a text</A>
|
|
<LI><A HREF="#toc6">PConj, phrasal conjunctions</A>
|
|
<LI><A HREF="#toc7">Voc, vocatives</A>
|
|
<LI><A HREF="#toc8">Utt, utterances</A>
|
|
</UL>
|
|
<LI><A HREF="#toc9">Auxiliary parameters for phrases and sentences</A>
|
|
<UL>
|
|
<LI><A HREF="#toc10">Pol, polarity</A>
|
|
<LI><A HREF="#toc11">Ant, anteriority</A>
|
|
<LI><A HREF="#toc12">Tense, tense</A>
|
|
<LI><A HREF="#toc13">ImpForm, imperative form</A>
|
|
</UL>
|
|
<LI><A HREF="#toc14">Sentences and clauses</A>
|
|
<UL>
|
|
<LI><A HREF="#toc15">S, sentences</A>
|
|
<LI><A HREF="#toc16">Cl, clauses</A>
|
|
</UL>
|
|
<LI><A HREF="#toc17">Verb phrases and imperatives</A>
|
|
<UL>
|
|
<LI><A HREF="#toc18">VP, verb phrases</A>
|
|
<LI><A HREF="#toc19">Imp, imperatives</A>
|
|
</UL>
|
|
<LI><A HREF="#toc20">Noun phrases and determiners</A>
|
|
<UL>
|
|
<LI><A HREF="#toc21">NP, noun phrases</A>
|
|
<LI><A HREF="#toc22">Det, determiners</A>
|
|
<LI><A HREF="#toc23">Quant, quantifiers with both sincular and plural forms</A>
|
|
<LI><A HREF="#toc24">QuantSg, singular quantifiers</A>
|
|
<LI><A HREF="#toc25">QuantPl, plural quantifiers</A>
|
|
<LI><A HREF="#toc26">Num, cardinal numerals</A>
|
|
<LI><A HREF="#toc27">Ord, ordinal numerals</A>
|
|
<LI><A HREF="#toc28">AdN, adnumerals</A>
|
|
<LI><A HREF="#toc29">Numeral, number words</A>
|
|
</UL>
|
|
<LI><A HREF="#toc30">Nouns</A>
|
|
<UL>
|
|
<LI><A HREF="#toc31">CN, common noun phrases</A>
|
|
</UL>
|
|
<LI><A HREF="#toc32">Adjectives and adverbs</A>
|
|
<UL>
|
|
<LI><A HREF="#toc33">AP, adjectival phrases</A>
|
|
<LI><A HREF="#toc34">Adv, adverbial phrases</A>
|
|
</UL>
|
|
<LI><A HREF="#toc35">Questions and relatives</A>
|
|
<UL>
|
|
<LI><A HREF="#toc36">QS, question sentences</A>
|
|
<LI><A HREF="#toc37">QCl, question clauses</A>
|
|
<LI><A HREF="#toc38">IP, interrogative pronouns</A>
|
|
<LI><A HREF="#toc39">IAdv, interrogative adverbs.</A>
|
|
<LI><A HREF="#toc40">RS, relative sentences</A>
|
|
<LI><A HREF="#toc41">RCl, relative clauses</A>
|
|
<LI><A HREF="#toc42">RP, relative pronouns</A>
|
|
<LI><A HREF="#toc43">Slash, objectless sentences</A>
|
|
</UL>
|
|
<LI><A HREF="#toc44">Lists for coordination</A>
|
|
<UL>
|
|
<LI><A HREF="#toc45">ListS, sentence lists</A>
|
|
<LI><A HREF="#toc46">ListAdv, adverb lists</A>
|
|
<LI><A HREF="#toc47">ListAP, adjectival phrase lists</A>
|
|
<LI><A HREF="#toc48">ListNP, noun phrase lists</A>
|
|
</UL>
|
|
</UL>
|
|
|
|
<P></P>
|
|
<HR NOSHADE SIZE=1>
|
|
<P></P>
|
|
<P>
|
|
Produced by
|
|
gfdoc - a rudimentary GF document generator.
|
|
(c) Aarne Ranta (<A HREF="mailto:aarne@cs.chalmers.se">aarne@cs.chalmers.se</A>) 2002 under GNU GPL.
|
|
</P>
|
|
<PRE>
|
|
incomplete resource Constructors = open Grammar in {
|
|
|
|
flags optimize=noexpand ;
|
|
</PRE>
|
|
<P></P>
|
|
<P>
|
|
This module gives access to the syntactic constructions of the
|
|
GF Resource Grammar library. Its main principle is simple:
|
|
to construct an object of type <CODE>C</CODE>, use the function <CODE>mkC</CODE>.
|
|
</P>
|
|
<P>
|
|
For example, an object of type <CODE>S</CODE> corresponding to the string
|
|
</P>
|
|
<P>
|
|
<CODE>John loves Mary</CODE>
|
|
</P>
|
|
<P>
|
|
is written
|
|
</P>
|
|
<P>
|
|
<CODE>mkS (mkCl (mkNP (mkPN "John")) (mkV2 "love") (mkNP (mkPN "Mary")))</CODE>
|
|
</P>
|
|
<P>
|
|
This module defines the syntactic constructors, which take trees as arguments.
|
|
Lexical constructors, which take strings as arguments, are defined in the
|
|
<CODE>Paradigms</CODE> modules separately for each language.
|
|
</P>
|
|
<P>
|
|
The recommended usage of this module is via the wrapper module <CODE>Syntax</CODE>,
|
|
which also contains the <CODE>Structural</CODE> (structural words).
|
|
Together with <CODE>Paradigms</CODE>, <CODE>Syntax</CODE> gives everything that is needed
|
|
to implement the concrete syntax for a langauge.
|
|
</P>
|
|
<A NAME="toc1"></A>
|
|
<H2>Principles of organization</H2>
|
|
<P>
|
|
To make the library easier to grasp and navigate, we have followed
|
|
a set of principles when organizing it:
|
|
</P>
|
|
<OL>
|
|
<LI>Each category <CODE>C</CODE> has an overloaded constructor <CODE>mkC</CODE>, with value type <CODE>C</CODE>.
|
|
<LI>With <CODE>mkC</CODE>, it is possible to construct any tree of type <CODE>C</CODE>, except
|
|
atomic ones, i.e. those that take no arguments, and
|
|
those whose argument types are exactly the same as in some other instance
|
|
<LI>To achieve completeness, the library therefore also has
|
|
for each atomic tree of type <CODE>C</CODE>, a constant suffixed <CODE>C</CODE>, and,
|
|
for other missing constructions, some operation suffixed <CODE>C</CODE>.
|
|
These constructors are listed immediately after the <CODE>mkC</CODE> group.
|
|
<LI>Those atomic constructors that are given in <CODE>Structural</CODE> are not repeated here.
|
|
<LI>In addition to the minimally complete set of constructions, many <CODE>mkC</CODE> groups
|
|
include some frequently needed special cases, with two possible logics:
|
|
default value (to decrease the number of arguments), and
|
|
direct arguments of an intervening constructor (to flatten the terms).
|
|
<LI>If such a special case is applied to some category in some rule, it is
|
|
also applied to all other rules in which the category appears.
|
|
<LI>The constructors in a group are listed, roughly,
|
|
<B>from the most common to the most general</B>. This does not of course specify
|
|
a total order.
|
|
<LI>Optional argument types are marked in parentheses. Although parentheses make no
|
|
difference in the way the GF compiler treats the types, their presence indicates
|
|
to the reader that the corresponding arguments can be left out; internally, the
|
|
library has an overload case for each such combination.
|
|
<LI>Each constructor case is equipped with an example that is built by that
|
|
case but could not be built with any other one.
|
|
</OL>
|
|
|
|
<A NAME="toc2"></A>
|
|
<H2>Texts, phrases, and utterances</H2>
|
|
<A NAME="toc3"></A>
|
|
<H3>Text: texts</H3>
|
|
<P>
|
|
A text is a list of phrases separated by punctuation marks.
|
|
The default punctuation mark is the full stop, and the default
|
|
continuation of a text is empty.
|
|
</P>
|
|
<PRE>
|
|
oper
|
|
mkText : overload {
|
|
mkText : Phr -> Text ; -- 1. But John walks.
|
|
mkText : Phr -> (Punct) -> (Text) -> Text ; -- 2. John walks? Yes.
|
|
</PRE>
|
|
<P></P>
|
|
<P>
|
|
A text can also be directly built from utterances, which in turn can
|
|
be directly built from sentences, present-tense clauses, questions, or
|
|
positive imperatives.
|
|
</P>
|
|
<PRE>
|
|
mkText : Utt -> Text ; -- 3. John.
|
|
mkText : S -> Text ; -- 4. John walked.
|
|
mkText : Cl -> Text ; -- 5. John walks.
|
|
mkText : QS -> Text ; -- 6. Did John walk?
|
|
mkText : Imp -> Text -- 7. Walk!
|
|
} ;
|
|
</PRE>
|
|
<P></P>
|
|
<P>
|
|
A text can also be empty.
|
|
</P>
|
|
<PRE>
|
|
emptyText : Text ; -- 8. (empty text)
|
|
</PRE>
|
|
<P></P>
|
|
<A NAME="toc4"></A>
|
|
<H3>Punct: punctuation marks</H3>
|
|
<P>
|
|
There are three punctuation marks that can separate phrases in a text.
|
|
</P>
|
|
<PRE>
|
|
fullStopPunct : Punct ; -- .
|
|
questMarkPunct : Punct ; -- ?
|
|
exclMarkPunct : Punct ; -- !
|
|
</PRE>
|
|
<P></P>
|
|
<A NAME="toc5"></A>
|
|
<H3>Phr: phrases in a text</H3>
|
|
<P>
|
|
Phrases are built from utterances by adding a phrasal conjunction
|
|
and a vocative, both of which are by default empty.
|
|
</P>
|
|
<PRE>
|
|
mkPhr : overload {
|
|
mkPhr : Utt -> Phr ; -- 1. why
|
|
mkPhr : (PConj) -> Utt -> (Voc) -> Phr ; -- 2. but why John
|
|
</PRE>
|
|
<P></P>
|
|
<P>
|
|
A phrase can also be directly built by a sentence, a present-tense
|
|
clause, a question, or a positive singular imperative.
|
|
</P>
|
|
<PRE>
|
|
mkPhr : S -> Phr ; -- 3. John walked
|
|
mkPhr : Cl -> Phr ; -- 4. John walks
|
|
mkPhr : QS -> Phr ; -- 5. did John walk
|
|
mkPhr : Imp -> Phr -- 6. walk
|
|
} ;
|
|
</PRE>
|
|
<P></P>
|
|
<A NAME="toc6"></A>
|
|
<H3>PConj, phrasal conjunctions</H3>
|
|
<P>
|
|
Any conjunction can be used as a phrasal conjunction.
|
|
More phrasal conjunctions are defined in <CODE>Structural</CODE>.
|
|
</P>
|
|
<PRE>
|
|
mkPConj : Conj -> PConj ; -- 1. and
|
|
</PRE>
|
|
<P></P>
|
|
<A NAME="toc7"></A>
|
|
<H3>Voc, vocatives</H3>
|
|
<P>
|
|
Any noun phrase can be turned into a vocative.
|
|
More vocatives are defined in <CODE>Structural</CODE>.
|
|
</P>
|
|
<PRE>
|
|
mkVoc : NP -> Voc ; -- 1. John
|
|
</PRE>
|
|
<P></P>
|
|
<A NAME="toc8"></A>
|
|
<H3>Utt, utterances</H3>
|
|
<P>
|
|
Utterances are formed from sentences, clauses, questions, and positive singular imperatives.
|
|
</P>
|
|
<PRE>
|
|
mkUtt : overload {
|
|
mkUtt : S -> Utt ; -- 1. John walked
|
|
mkUtt : Cl -> Utt ; -- 2. John walks
|
|
mkUtt : QS -> Utt ; -- 3. did John walk
|
|
mkUtt : Imp -> Utt ; -- 4. love yourself
|
|
</PRE>
|
|
<P></P>
|
|
<P>
|
|
Imperatives can also vary in <CODE>ImpForm</CODE> (number/politeness) and
|
|
polarity.
|
|
</P>
|
|
<PRE>
|
|
mkUtt : (ImpForm) -> (Pol) -> Imp -> Utt ; -- 5. don't love yourselves
|
|
</PRE>
|
|
<P></P>
|
|
<P>
|
|
Utterances can also be formed from interrogative phrases and
|
|
interrogative adverbials, noun phrases, adverbs, and verb phrases.
|
|
</P>
|
|
<PRE>
|
|
mkUtt : IP -> Utt ; -- 6. who
|
|
mkUtt : IAdv -> Utt ; -- 7. why
|
|
mkUtt : NP -> Utt ; -- 8. John
|
|
mkUtt : Adv -> Utt ; -- 9. here
|
|
mkUtt : VP -> Utt -- 10. to walk
|
|
} ;
|
|
</PRE>
|
|
<P></P>
|
|
<P>
|
|
The plural first-person imperative is a special construction.
|
|
</P>
|
|
<PRE>
|
|
lets_Utt : VP -> Utt ; -- 11. let's walk
|
|
</PRE>
|
|
<P></P>
|
|
<A NAME="toc9"></A>
|
|
<H2>Auxiliary parameters for phrases and sentences</H2>
|
|
<A NAME="toc10"></A>
|
|
<H3>Pol, polarity</H3>
|
|
<P>
|
|
Polarity is a parameter that sets a clause to positive or negative
|
|
form. Since positive is the default, it need never be given explicitly.
|
|
</P>
|
|
<PRE>
|
|
positivePol : Pol ; -- (John walks) [default]
|
|
negativePol : Pol ; -- (John doesn't walk)
|
|
</PRE>
|
|
<P></P>
|
|
<A NAME="toc11"></A>
|
|
<H3>Ant, anteriority</H3>
|
|
<P>
|
|
Anteriority is a parameter that presents an event as simultaneous or
|
|
anterior to some other reference time.
|
|
Since simultaneous is the default, it need never be given explicitly.
|
|
</P>
|
|
<PRE>
|
|
simultaneousAnt : Ant ; -- (John walks) [default]
|
|
anteriorAnt : Ant ; -- (John has walked) --# notpresent
|
|
</PRE>
|
|
<P></P>
|
|
<A NAME="toc12"></A>
|
|
<H3>Tense, tense</H3>
|
|
<P>
|
|
Tense is a parameter that relates the time of an event
|
|
to the time of speaking about it.
|
|
Since present is the default, it need never be given explicitly.
|
|
</P>
|
|
<PRE>
|
|
presentTense : Tense ; -- (John walks) [default]
|
|
pastTense : Tense ; -- (John walked) --# notpresent
|
|
futureTense : Tense ; -- (John will walk) --# notpresent
|
|
conditionalTense : Tense ; -- (John would walk) --# notpresent
|
|
</PRE>
|
|
<P></P>
|
|
<A NAME="toc13"></A>
|
|
<H3>ImpForm, imperative form</H3>
|
|
<P>
|
|
Imperative form is a parameter that sets the form of imperative
|
|
by reference to the person or persons addressed.
|
|
Since singular is the default, it need never be given explicitly.
|
|
</P>
|
|
<PRE>
|
|
singularImpForm : ImpForm ; -- (help yourself) [default]
|
|
pluralImpForm : ImpForm ; -- (help yourselves)
|
|
politeImpForm : ImpForm ; -- (help yourself) (polite singular)
|
|
</PRE>
|
|
<P></P>
|
|
<A NAME="toc14"></A>
|
|
<H2>Sentences and clauses</H2>
|
|
<A NAME="toc15"></A>
|
|
<H3>S, sentences</H3>
|
|
<P>
|
|
A sentence has a fixed tense, anteriority and polarity.
|
|
</P>
|
|
<PRE>
|
|
mkS : overload {
|
|
mkS : Cl -> S ; -- 1. John walks
|
|
mkS : (Tense) -> (Ant) -> (Pol) -> Cl -> S ; -- 2. John wouldn't have walked
|
|
</PRE>
|
|
<P></P>
|
|
<P>
|
|
Sentences can be combined with conjunctions. This can apply to a pair
|
|
of sentences, but also to a list of more than two.
|
|
</P>
|
|
<PRE>
|
|
mkS : Conj -> S -> S -> S ; -- 3. John walks and I run
|
|
mkS : Conj -> ListS -> S ; -- 4. John walks, I run and you sleep
|
|
mkS : DConj -> S -> S -> S ; -- 5. either John walk or I run
|
|
mkS : DConj -> ListS -> S ; -- 6. either John walks, I run or you sleep
|
|
</PRE>
|
|
<P></P>
|
|
<P>
|
|
A sentence can be prefixed by an adverb.
|
|
</P>
|
|
<PRE>
|
|
mkS : Adv -> S -> S -- 7. today, John walks
|
|
} ;
|
|
</PRE>
|
|
<P></P>
|
|
<A NAME="toc16"></A>
|
|
<H3>Cl, clauses</H3>
|
|
<P>
|
|
A clause has a variable tense, anteriority and polarity.
|
|
A clause can be built from a subject noun phrase
|
|
with a verb and appropriate arguments.
|
|
</P>
|
|
<PRE>
|
|
mkCl : overload {
|
|
mkCl : NP -> V -> Cl ; -- 1. John walks
|
|
mkCl : NP -> V2 -> NP -> Cl ; -- 2. John loves her
|
|
mkCl : NP -> V3 -> NP -> NP -> Cl ; -- 3. John sends it to her
|
|
mkCl : NP -> VV -> VP -> Cl ; -- 4. John wants to walk
|
|
mkCl : NP -> VS -> S -> Cl ; -- 5. John says that it is good
|
|
mkCl : NP -> VQ -> QS -> Cl ; -- 6. John wonders if it is good
|
|
mkCl : NP -> VA -> AP -> Cl ; -- 7. John becomes old
|
|
mkCl : NP -> V2A ->NP -> AP -> Cl ; -- 8. John paints it red
|
|
mkCl : NP -> A -> Cl ; -- 9. John is old
|
|
mkCl : NP -> A -> NP -> Cl ; -- 10. John is older than her
|
|
mkCl : NP -> A2 -> NP -> Cl ; -- 11. John is married to her
|
|
mkCl : NP -> AP -> Cl ; -- 12. John is very old
|
|
mkCl : NP -> N -> Cl ; -- 13. John is a man
|
|
mkCl : NP -> CN -> Cl ; -- 14. John is an old man
|
|
mkCl : NP -> NP -> Cl ; -- 15. John is the man
|
|
mkCl : NP -> Adv -> Cl ; -- 16. John is here
|
|
</PRE>
|
|
<P></P>
|
|
<P>
|
|
As the general rule, a clause can be built from a subject noun phrase and
|
|
a verb phrase.
|
|
</P>
|
|
<PRE>
|
|
mkCl : NP -> VP -> Cl ; -- 17. John walks here
|
|
</PRE>
|
|
<P></P>
|
|
<P>
|
|
Subjectless verb phrases are used for impersonal actions.
|
|
</P>
|
|
<PRE>
|
|
mkCl : V -> Cl ; -- 18. it rains
|
|
mkCl : VP -> Cl ; -- 19. it is raining
|
|
</PRE>
|
|
<P></P>
|
|
<P>
|
|
Existentials are a special form of clauses.
|
|
</P>
|
|
<PRE>
|
|
mkCl : N -> Cl ; -- 20. there is a house
|
|
mkCl : CN -> Cl ; -- 21. there is an old houses
|
|
mkCl : NP -> Cl ; -- 22. there are five houses
|
|
</PRE>
|
|
<P></P>
|
|
<P>
|
|
There are also special forms in which a noun phrase or an adverb is
|
|
emphasized.
|
|
</P>
|
|
<PRE>
|
|
mkCl : NP -> RS -> Cl ; -- 23. it is John that walks
|
|
mkCl : Adv -> S -> Cl -- 24. it is here John walks
|
|
} ;
|
|
</PRE>
|
|
<P></P>
|
|
<P>
|
|
Generic clauses are one with an impersonal subject.
|
|
</P>
|
|
<PRE>
|
|
genericCl : VP -> Cl ; -- 25. one walks
|
|
</PRE>
|
|
<P></P>
|
|
<A NAME="toc17"></A>
|
|
<H2>Verb phrases and imperatives</H2>
|
|
<A NAME="toc18"></A>
|
|
<H3>VP, verb phrases</H3>
|
|
<P>
|
|
A verb phrase is formed from a verb with appropriate arguments.
|
|
</P>
|
|
<PRE>
|
|
mkVP : overload {
|
|
mkVP : V -> VP ; -- 1. walk
|
|
mkVP : V2 -> NP -> VP ; -- 2. love her
|
|
mkVP : V3 -> NP -> NP -> VP ; -- 3. send it to her
|
|
mkVP : VV -> VP -> VP ; -- 4. want to walk
|
|
mkVP : VS -> S -> VP ; -- 5. know that she walks
|
|
mkVP : VQ -> QS -> VP ; -- 6. ask if she walks
|
|
mkVP : VA -> AP -> VP ; -- 7. become old
|
|
mkVP : V2A -> NP -> AP -> VP ; -- 8. paint it red
|
|
</PRE>
|
|
<P></P>
|
|
<P>
|
|
The verb can also be a copula (<I>be</I>), and the relevant argument is
|
|
then the complement adjective or noun phrase.
|
|
</P>
|
|
<PRE>
|
|
mkVP : A -> VP ; -- 9. be warm
|
|
mkVP : AP -> VP ; -- 12. be very warm
|
|
mkVP : A -> NP -> VP ; -- 10. be older than her
|
|
mkVP : A2 -> NP -> VP ; -- 11. be married to her
|
|
mkVP : N -> VP ; -- 13. be a man
|
|
mkVP : CN -> VP ; -- 14. be an old man
|
|
mkVP : NP -> VP ; -- 15. be the man
|
|
mkVP : Adv -> VP ; -- 16. be here
|
|
</PRE>
|
|
<P></P>
|
|
<P>
|
|
A verb phrase can be modified with a postverbal or a preverbal adverb.
|
|
</P>
|
|
<PRE>
|
|
mkVP : VP -> Adv -> VP ; -- 17. sleep here
|
|
mkVP : AdV -> VP -> VP -- 18. always sleep
|
|
} ;
|
|
</PRE>
|
|
<P></P>
|
|
<P>
|
|
Two-place verbs can be used reflexively.
|
|
</P>
|
|
<PRE>
|
|
reflexiveVP : V2 -> VP ; -- 19. love itself
|
|
</PRE>
|
|
<P></P>
|
|
<P>
|
|
Two-place verbs can also be used in the passive, with or without an agent.
|
|
</P>
|
|
<PRE>
|
|
passiveVP : overload {
|
|
passiveVP : V2 -> VP ; -- 20. be loved
|
|
passiveVP : V2 -> NP -> VP ; -- 21. be loved by her
|
|
} ;
|
|
</PRE>
|
|
<P></P>
|
|
<P>
|
|
A verb phrase can be turned into the progressive form.
|
|
</P>
|
|
<PRE>
|
|
progressiveVP : VP -> VP ; -- 22. be sleeping
|
|
</PRE>
|
|
<P></P>
|
|
<A NAME="toc19"></A>
|
|
<H3>Imp, imperatives</H3>
|
|
<P>
|
|
Imperatives are formed from verbs and their arguments; as the general
|
|
rule, from verb phrases.
|
|
</P>
|
|
<PRE>
|
|
mkImp : overload {
|
|
mkImp : V -> Imp ; -- go
|
|
mkImp : V2 -> NP -> Imp ; -- take it
|
|
mkImp : VP -> Imp -- go there now
|
|
} ;
|
|
</PRE>
|
|
<P></P>
|
|
<A NAME="toc20"></A>
|
|
<H2>Noun phrases and determiners</H2>
|
|
<A NAME="toc21"></A>
|
|
<H3>NP, noun phrases</H3>
|
|
<P>
|
|
A noun phrases can be built from a determiner and a common noun (<CODE>CN</CODE>) .
|
|
For determiners, the special cases of quantifiers, numerals, integers,
|
|
and possessive pronouns are provided. For common nouns, the
|
|
special case of a simple common noun (<CODE>N</CODE>) is always provided.
|
|
</P>
|
|
<PRE>
|
|
mkNP : overload {
|
|
mkNP : Det -> N -> NP ; -- 1. the first man
|
|
mkNP : Det -> CN -> NP ; -- 2. the first old man
|
|
mkNP : QuantSg -> N -> NP ; -- 3. this man
|
|
mkNP : QuantSg -> CN -> NP ; -- 4. this old man
|
|
mkNP : QuantPl -> N -> NP ; -- 5. these men
|
|
mkNP : QuantPl -> CN -> NP ; -- 6. these old men
|
|
mkNP : Numeral -> N -> NP ; -- 7. twenty men
|
|
mkNP : Numeral -> CN -> NP ; -- 8. twenty old men
|
|
mkNP : Int -> N -> NP ; -- 9. 45 men
|
|
mkNP : Int -> CN -> NP ; -- 10. 45 old men
|
|
mkNP : Num -> N -> NP ; -- 11. almost twenty men
|
|
mkNP : Num -> CN -> NP ; -- 12. almost twenty old men
|
|
mkNP : Pron -> N -> NP ; -- 13. my man
|
|
mkNP : Pron -> CN -> NP; -- 14. my old man
|
|
</PRE>
|
|
<P></P>
|
|
<P>
|
|
Proper names and pronouns can be used as noun phrases.
|
|
</P>
|
|
<PRE>
|
|
mkNP : PN -> NP ; -- 15. John
|
|
mkNP : Pron -> NP ; -- 16. he
|
|
</PRE>
|
|
<P></P>
|
|
<P>
|
|
A noun phrase once formed can be prefixed by a predeterminer and
|
|
suffixed by a past participle or an adverb.
|
|
</P>
|
|
<PRE>
|
|
mkNP : Predet -> NP -> NP ; -- 17. only John
|
|
mkNP : NP -> V2 -> NP ; -- 18. John killed
|
|
mkNP : NP -> Adv -> NP ; -- 19. John in Paris
|
|
</PRE>
|
|
<P></P>
|
|
<P>
|
|
A conjunction can be formed both from two noun phrases and a longer
|
|
list of them.
|
|
</P>
|
|
<PRE>
|
|
mkNP : Conj -> NP -> NP -> NP ; -- 20. John and I
|
|
mkNP : Conj -> ListNP -> NP ; -- 21. John, I, and that
|
|
mkNP : DConj -> NP -> NP -> NP ; -- 22. either John or I
|
|
mkNP : DConj -> ListNP -> NP -- 23. either John, I, or that
|
|
|
|
} ;
|
|
</PRE>
|
|
<P></P>
|
|
<A NAME="toc22"></A>
|
|
<H3>Det, determiners</H3>
|
|
<P>
|
|
A determiner is either a singular or a plural one.
|
|
Both have a quantifier and an optional ordinal; the plural
|
|
determiner also has an optional numeral.
|
|
</P>
|
|
<PRE>
|
|
mkDet : overload {
|
|
mkDet : QuantSg -> Det ; -- 1. this
|
|
mkDet : QuantSg -> (Ord) -> Det ; -- 2. this first
|
|
mkDet : QuantPl -> Det ; -- 3. these
|
|
mkDet : QuantPl -> (Num) -> (Ord) -> Det ; -- 4. these five best
|
|
</PRE>
|
|
<P></P>
|
|
<P>
|
|
Quantifiers that have both singular and plural forms are by default used as
|
|
singular determiners. If a numeral is added, the plural form is chosen.
|
|
</P>
|
|
<PRE>
|
|
mkDet : Quant -> Det ; -- 5. this
|
|
mkDet : Quant -> Num -> Det ; -- 6. these five
|
|
</PRE>
|
|
<P></P>
|
|
<P>
|
|
Numerals, their special cases integers and digits, and possessive pronouns can be
|
|
used as determiners.
|
|
</P>
|
|
<PRE>
|
|
mkDet : Num -> Det ; -- 7. almost twenty
|
|
mkDet : Numeral -> Det ; -- 8. five
|
|
mkDet : Int -> Det ; -- 9. 51
|
|
mkDet : Pron -> Det -- 10. my
|
|
} ;
|
|
</PRE>
|
|
<P></P>
|
|
<P>
|
|
The definite and indefinite articles are commonly used determiners.
|
|
</P>
|
|
<PRE>
|
|
defSgDet : Det ; -- 11. the (house)
|
|
defPlDet : Det ; -- 12. the (houses)
|
|
indefSgDet : Det ; -- 13. a (house)
|
|
indefPlDet : Det ; -- 14. (houses)
|
|
</PRE>
|
|
<P></P>
|
|
<A NAME="toc23"></A>
|
|
<H3>Quant, quantifiers with both sincular and plural forms</H3>
|
|
<P>
|
|
Definite and indefinite articles have both singular and plural
|
|
forms (even though the
|
|
plural indefinite is empty in most languages).
|
|
</P>
|
|
<PRE>
|
|
defQuant : Quant ; -- 1. the
|
|
indefQuant : Quant ; -- 2. a
|
|
</PRE>
|
|
<P></P>
|
|
<P>
|
|
More quantifiers are available in the <CODE>Structural</CODE> module.
|
|
</P>
|
|
<A NAME="toc24"></A>
|
|
<H3>QuantSg, singular quantifiers</H3>
|
|
<P>
|
|
From quantifiers that can have both forms, this constructor
|
|
builds the singular form.
|
|
</P>
|
|
<PRE>
|
|
mkQuantSg : Quant -> QuantSg ; -- 1. this
|
|
</PRE>
|
|
<P></P>
|
|
<P>
|
|
The mass noun phrase constructor is treated as a singular quantifier.
|
|
</P>
|
|
<PRE>
|
|
massQuant : QuantSg ; -- 2. (mass terms)
|
|
</PRE>
|
|
<P></P>
|
|
<P>
|
|
More singular quantifiers are available in the <CODE>Structural</CODE> module.
|
|
</P>
|
|
<A NAME="toc25"></A>
|
|
<H3>QuantPl, plural quantifiers</H3>
|
|
<P>
|
|
From quantifiers that can have both forms, this constructor
|
|
builds the plural form.
|
|
</P>
|
|
<PRE>
|
|
mkQuantPl : Quant -> QuantPl ; -- 1. these
|
|
</PRE>
|
|
<P></P>
|
|
<P>
|
|
More plural quantifiers are available in the <CODE>Structural</CODE> module.
|
|
</P>
|
|
<A NAME="toc26"></A>
|
|
<H3>Num, cardinal numerals</H3>
|
|
<P>
|
|
Numerals can be formed from number words (<CODE>Numeral</CODE>), their special case digits,
|
|
and from symbolic integers.
|
|
</P>
|
|
<PRE>
|
|
mkNum : overload {
|
|
mkNum : Numeral -> Num ; -- 1. twenty
|
|
mkNum : Int -> Num ; -- 2. 51
|
|
</PRE>
|
|
<P></P>
|
|
<P>
|
|
A numeral can be modified by an adnumeral.
|
|
</P>
|
|
<PRE>
|
|
mkNum : AdN -> Num -> Num -- 3. almost ten
|
|
} ;
|
|
</PRE>
|
|
<P></P>
|
|
<A NAME="toc27"></A>
|
|
<H3>Ord, ordinal numerals</H3>
|
|
<P>
|
|
Just like cardinals, ordinals can be formed from number words (<CODE>Numeral</CODE>)
|
|
and from symbolic integers.
|
|
</P>
|
|
<PRE>
|
|
mkOrd : overload {
|
|
mkOrd : Numeral -> Ord ; -- 1. twentieth
|
|
mkOrd : Int -> Ord ; -- 2. 51st
|
|
</PRE>
|
|
<P></P>
|
|
<P>
|
|
Also adjectives in the superlative form can appear on ordinal positions.
|
|
</P>
|
|
<PRE>
|
|
mkOrd : A -> Ord -- 3. best
|
|
} ;
|
|
</PRE>
|
|
<P></P>
|
|
<A NAME="toc28"></A>
|
|
<H3>AdN, adnumerals</H3>
|
|
<P>
|
|
Comparison adverbs can be used as adnumerals.
|
|
</P>
|
|
<PRE>
|
|
mkAdN : CAdv -> AdN ; -- 1. more than
|
|
</PRE>
|
|
<P></P>
|
|
<A NAME="toc29"></A>
|
|
<H3>Numeral, number words</H3>
|
|
<P>
|
|
Digits and some <I>round</I> numbers are here given as shorthands.
|
|
</P>
|
|
<PRE>
|
|
n1_Numeral : Numeral ; -- 1. one
|
|
n2_Numeral : Numeral ; -- 2. two
|
|
n3_Numeral : Numeral ; -- 3. three
|
|
n4_Numeral : Numeral ; -- 4. four
|
|
n5_Numeral : Numeral ; -- 5. five
|
|
n6_Numeral : Numeral ; -- 6. six
|
|
n7_Numeral : Numeral ; -- 7. seven
|
|
n8_Numeral : Numeral ; -- 8. eight
|
|
n9_Numeral : Numeral ; -- 9. nine
|
|
n10_Numeral : Numeral ; -- 10. ten
|
|
n20_Numeral : Numeral ; -- 11. twenty
|
|
n100_Numeral : Numeral ; -- 12. hundred
|
|
n1000_Numeral : Numeral ; -- 13. thousand
|
|
</PRE>
|
|
<P></P>
|
|
<P>
|
|
See <CODE>Numeral</CODE> for the full set of constructors, or use <CODE>Int</CODE> for other numbers.
|
|
</P>
|
|
<A NAME="toc30"></A>
|
|
<H2>Nouns</H2>
|
|
<A NAME="toc31"></A>
|
|
<H3>CN, common noun phrases</H3>
|
|
<PRE>
|
|
mkCN : overload {
|
|
</PRE>
|
|
<P></P>
|
|
<P>
|
|
The most frequent way of forming common noun phrases is from atomic nouns <CODE>N</CODE>.
|
|
</P>
|
|
<PRE>
|
|
mkCN : N -> CN ; -- 1. house
|
|
</PRE>
|
|
<P></P>
|
|
<P>
|
|
Common noun phrases can be formed from relational nouns by providing arguments.
|
|
</P>
|
|
<PRE>
|
|
mkCN : N2 -> NP -> CN ; -- 2. mother of John
|
|
mkCN : N3 -> NP -> NP -> CN ; -- 3. distance from this city to Paris
|
|
</PRE>
|
|
<P></P>
|
|
<P>
|
|
Relational nouns can also be used without their arguments.
|
|
</P>
|
|
<PRE>
|
|
mkCN : N2 -> CN ; -- 4. son
|
|
mkCN : N3 -> CN ; -- 5. flight
|
|
</PRE>
|
|
<P></P>
|
|
<P>
|
|
A common noun phrase can be modified by adjectival phrase. We give special
|
|
cases of this, where one or both of the arguments are atomic.
|
|
</P>
|
|
<PRE>
|
|
mkCN : A -> N -> CN ; -- 6. big house
|
|
mkCN : A -> CN -> CN ; -- 7. big blue house
|
|
mkCN : AP -> N -> CN ; -- 8. very big house
|
|
mkCN : AP -> CN -> CN ; -- 9. very big blue house
|
|
</PRE>
|
|
<P></P>
|
|
<P>
|
|
A common noun phrase can be modified by a relative clause or an adverb.
|
|
</P>
|
|
<PRE>
|
|
mkCN : N -> RS -> CN ; -- 10. house that John loves
|
|
mkCN : CN -> RS -> CN ; -- 11. big house that John loves
|
|
mkCN : N -> Adv -> CN ; -- 12. house in the city
|
|
mkCN : CN -> Adv -> CN ; -- 13. big house in the city
|
|
</PRE>
|
|
<P></P>
|
|
<P>
|
|
For some nouns it makes sense to modify them by sentences,
|
|
questions, or infinitives. But syntactically this is possible for
|
|
all nouns.
|
|
</P>
|
|
<PRE>
|
|
mkCN : CN -> S -> CN ; -- 14. rule that John walks
|
|
mkCN : CN -> QS -> CN ; -- 15. question if John walks
|
|
mkCN : CN -> VP -> CN ; -- 16. reason to walk
|
|
</PRE>
|
|
<P></P>
|
|
<P>
|
|
A noun can be used in apposition to a noun phrase, especially a proper name.
|
|
</P>
|
|
<PRE>
|
|
mkCN : N -> NP -> CN ; -- 17. king John
|
|
mkCN : CN -> NP -> CN -- 18. old king John
|
|
} ;
|
|
</PRE>
|
|
<P></P>
|
|
<A NAME="toc32"></A>
|
|
<H2>Adjectives and adverbs</H2>
|
|
<A NAME="toc33"></A>
|
|
<H3>AP, adjectival phrases</H3>
|
|
<PRE>
|
|
mkAP : overload {
|
|
</PRE>
|
|
<P></P>
|
|
<P>
|
|
Adjectival phrases can be formed from atomic adjectives by using the positive form or
|
|
the comparative with a complement
|
|
</P>
|
|
<PRE>
|
|
mkAP : A -> AP ; -- 1. old
|
|
mkAP : A -> NP -> AP ; -- 2. older than John
|
|
</PRE>
|
|
<P></P>
|
|
<P>
|
|
Relational adjectives can be used with a complement or a reflexive
|
|
</P>
|
|
<PRE>
|
|
mkAP : A2 -> NP -> AP ; -- 3. married to her
|
|
mkAP : A2 -> AP ; -- 4. married to myself
|
|
</PRE>
|
|
<P></P>
|
|
<P>
|
|
Some adjectival phrases can take as complements sentences,
|
|
questions, or infinitives. Syntactically this is possible for
|
|
all adjectives.
|
|
</P>
|
|
<PRE>
|
|
mkAP : AP -> S -> AP ; -- 5. probable that John walks
|
|
mkAP : AP -> QS -> AP ; -- 6. uncertain if John walks
|
|
mkAP : AP -> VP -> AP ; -- 7. ready to go
|
|
</PRE>
|
|
<P></P>
|
|
<P>
|
|
An adjectival phrase can be modified by an adadjective.
|
|
</P>
|
|
<PRE>
|
|
mkAP : AdA -> A -> AP ; -- 8. very old
|
|
mkAP : AdA -> AP -> AP ; -- 9. very very old
|
|
</PRE>
|
|
<P></P>
|
|
<P>
|
|
Conjunction can be formed from two or more adjectival phrases.
|
|
</P>
|
|
<PRE>
|
|
mkAP : Conj -> AP -> AP -> AP ; -- 10. old and big
|
|
mkAP : Conj -> ListAP -> AP ; -- 11. old, big, and warm
|
|
mkAP : DConj -> AP -> AP -> AP ; -- 12. either old or big
|
|
mkAP : DConj -> ListAP -> AP -- 13. either old, big, or warm
|
|
|
|
} ;
|
|
</PRE>
|
|
<P></P>
|
|
<A NAME="toc34"></A>
|
|
<H3>Adv, adverbial phrases</H3>
|
|
<PRE>
|
|
mkAdv : overload {
|
|
</PRE>
|
|
<P></P>
|
|
<P>
|
|
Adverbs can be formed from adjectives.
|
|
</P>
|
|
<PRE>
|
|
mkAdv : A -> Adv ; -- 1. warmly
|
|
</PRE>
|
|
<P></P>
|
|
<P>
|
|
Prepositional phrases are treated as adverbs.
|
|
</P>
|
|
<PRE>
|
|
mkAdv : Prep -> NP -> Adv ; -- 2. with John
|
|
</PRE>
|
|
<P></P>
|
|
<P>
|
|
Subordinate sentences are treated as adverbs.
|
|
</P>
|
|
<PRE>
|
|
mkAdv : Subj -> S -> Adv ; -- 3. when John walks
|
|
</PRE>
|
|
<P></P>
|
|
<P>
|
|
An adjectival adverb can be compared to a noun phrase or a sentence.
|
|
</P>
|
|
<PRE>
|
|
mkAdv : CAdv -> A -> NP -> Adv ; -- 4. more warmly than John
|
|
mkAdv : CAdv -> A -> S -> Adv ; -- 5. more warmly than John walks
|
|
</PRE>
|
|
<P></P>
|
|
<P>
|
|
Adverbs can be modified by adadjectives.
|
|
</P>
|
|
<PRE>
|
|
mkAdv : AdA -> Adv -> Adv ; -- 6. very warmly
|
|
</PRE>
|
|
<P></P>
|
|
<P>
|
|
Conjunction can be formed from two or more adverbial phrases.
|
|
</P>
|
|
<PRE>
|
|
mkAdv : Conj -> Adv -> Adv -> Adv ; -- 7. here and now
|
|
mkAdv : Conj -> ListAdv -> Adv ; -- 8. with John, here and now
|
|
mkAdv : DConj -> Adv -> Adv -> Adv ; -- 9. either here or now
|
|
mkAdv : DConj -> ListAdv -> Adv -- 10. either here, now, or with John
|
|
} ;
|
|
</PRE>
|
|
<P></P>
|
|
<A NAME="toc35"></A>
|
|
<H2>Questions and relatives</H2>
|
|
<A NAME="toc36"></A>
|
|
<H3>QS, question sentences</H3>
|
|
<PRE>
|
|
mkQS : overload {
|
|
</PRE>
|
|
<P></P>
|
|
<P>
|
|
Just like a sentence <CODE>S</CODE> is built from a clause <CODE>Cl</CODE>,
|
|
a question sentence <CODE>QS</CODE> is built from
|
|
a question clause <CODE>QCl</CODE> by fixing tense, anteriority and polarity.
|
|
Any of these arguments can be omitted, which results in the
|
|
default (present, simultaneous, and positive, respectively).
|
|
</P>
|
|
<PRE>
|
|
mkQS : QCl -> QS ; -- 1. who walks
|
|
mkQS : (Tense) -> (Ant) -> (Pol) -> QCl -> QS ; -- 2. who wouldn't have walked
|
|
</PRE>
|
|
<P></P>
|
|
<P>
|
|
Since 'yes-no' question clauses can be built from clauses (see below),
|
|
we give a shortcut
|
|
for building a question sentence directly from a clause, using the defaults
|
|
present, simultaneous, and positive.
|
|
</P>
|
|
<PRE>
|
|
mkQS : Cl -> QS -- 3. does John walk
|
|
} ;
|
|
</PRE>
|
|
<P></P>
|
|
<A NAME="toc37"></A>
|
|
<H3>QCl, question clauses</H3>
|
|
<PRE>
|
|
mkQCl : overload {
|
|
</PRE>
|
|
<P></P>
|
|
<P>
|
|
'Yes-no' question clauses are built from 'declarative' clauses.
|
|
</P>
|
|
<PRE>
|
|
mkQCl : Cl -> QCl ; -- 1. does John walk
|
|
</PRE>
|
|
<P></P>
|
|
<P>
|
|
'Wh' questions are built from interrogative pronouns in subject
|
|
or object position. The former uses a verb phrase; we don't give
|
|
shortcuts for verb-argument sequences as we do for clauses.
|
|
The latter uses the 'slash' category of objectless clauses
|
|
(see below); we give the common special case with a two-place verb.
|
|
</P>
|
|
<PRE>
|
|
mkQCl : IP -> VP -> QCl ; -- 2. who walks
|
|
mkQCl : IP -> NP -> V2 -> QCl ; -- 3. whom does John love
|
|
mkQCl : IP -> Slash -> QCl ; -- 4. whom does John love today
|
|
</PRE>
|
|
<P></P>
|
|
<P>
|
|
Adverbial 'wh' questions are built with interrogative adverbials, with the
|
|
special case of prepositional phrases with interrogative pronouns.
|
|
</P>
|
|
<PRE>
|
|
mkQCl : IAdv -> Cl -> QCl ; -- 5. why does John walk
|
|
mkQCl : Prep -> IP -> Cl -> QCl ; -- 6. with who does John walk
|
|
</PRE>
|
|
<P></P>
|
|
<P>
|
|
An interrogative adverbial can serve as the complement of a copula.
|
|
</P>
|
|
<PRE>
|
|
mkQCl : IAdv -> NP -> QCl ; -- 7. where is John
|
|
</PRE>
|
|
<P></P>
|
|
<P>
|
|
Existentials are a special construction.
|
|
</P>
|
|
<PRE>
|
|
mkQCl : IP -> QCl -- 8. what is there
|
|
} ;
|
|
</PRE>
|
|
<P></P>
|
|
<A NAME="toc38"></A>
|
|
<H3>IP, interrogative pronouns</H3>
|
|
<PRE>
|
|
mkIP : overload {
|
|
</PRE>
|
|
<P></P>
|
|
<P>
|
|
Interrogative pronouns
|
|
can be formed much like noun phrases, by using interrogative determiners.
|
|
</P>
|
|
<PRE>
|
|
mkIP : IDet -> N -> IP ; -- 1. which city
|
|
mkIP : IDet -> (Num) -> (Ord) -> CN -> IP ; -- 2. which five best cities
|
|
</PRE>
|
|
<P></P>
|
|
<P>
|
|
An interrogative pronoun can be modified by an adverb.
|
|
</P>
|
|
<PRE>
|
|
mkIP : IP -> Adv -> IP -- 3. who in Paris
|
|
} ;
|
|
</PRE>
|
|
<P></P>
|
|
<P>
|
|
More interrogative pronouns and determiners can be found in <CODE>Structural</CODE>.
|
|
</P>
|
|
<A NAME="toc39"></A>
|
|
<H3>IAdv, interrogative adverbs.</H3>
|
|
<P>
|
|
In addition to the interrogative adverbs defined in the <CODE>Structural</CODE> lexicon, they
|
|
can be formed as prepositional phrases from interrogative pronouns.
|
|
</P>
|
|
<PRE>
|
|
mkIAdv : Prep -> IP -> IAdv ; -- 1. in which city
|
|
</PRE>
|
|
<P></P>
|
|
<P>
|
|
More interrogative adverbs are given in <CODE>Structural</CODE>.
|
|
</P>
|
|
<A NAME="toc40"></A>
|
|
<H3>RS, relative sentences</H3>
|
|
<P>
|
|
Just like a sentence <CODE>S</CODE> is built from a clause <CODE>Cl</CODE>,
|
|
a relative sentence <CODE>RS</CODE> is built from
|
|
a relative clause <CODE>RCl</CODE> by fixing the tense, anteriority and polarity.
|
|
Any of these arguments
|
|
can be omitted, which results in the default (present, simultaneous,
|
|
and positive, respectively).
|
|
</P>
|
|
<PRE>
|
|
mkRS : overload {
|
|
mkRS : RCl -> RS ; -- 1. that walk
|
|
mkRS : (Tense) -> (Ant) -> (Pol) -> RCl -> RS -- 2. that wouldn't have walked
|
|
} ;
|
|
</PRE>
|
|
<P></P>
|
|
<A NAME="toc41"></A>
|
|
<H3>RCl, relative clauses</H3>
|
|
<PRE>
|
|
mkRCl : overload {
|
|
</PRE>
|
|
<P></P>
|
|
<P>
|
|
Relative clauses are built from relative pronouns in subject or object position.
|
|
The former uses a verb phrase; we don't give
|
|
shortcuts for verb-argument sequences as we do for clauses.
|
|
The latter uses the 'slash' category of objectless clauses (see below);
|
|
we give the common special case with a two-place verb.
|
|
</P>
|
|
<PRE>
|
|
mkRCl : RP -> VP -> RCl ; -- 1. that walk
|
|
mkRCl : RP -> NP -> V2 -> RCl ; -- 2. which John loves
|
|
mkRCl : RP -> Slash -> RCl ; -- 3. which John loves today
|
|
</PRE>
|
|
<P></P>
|
|
<P>
|
|
There is a simple 'such that' construction for forming relative
|
|
clauses from clauses.
|
|
</P>
|
|
<PRE>
|
|
mkRCl : Cl -> RCl -- 4. such that John loves her
|
|
} ;
|
|
</PRE>
|
|
<P></P>
|
|
<A NAME="toc42"></A>
|
|
<H3>RP, relative pronouns</H3>
|
|
<P>
|
|
There is an atomic relative pronoun
|
|
</P>
|
|
<PRE>
|
|
which_RP : RP ; -- 1. which
|
|
</PRE>
|
|
<P></P>
|
|
<P>
|
|
A relative pronoun can be made into a kind of a prepositional phrase.
|
|
</P>
|
|
<PRE>
|
|
mkRP : Prep -> NP -> RP -> RP ; -- 2. all the houses in which
|
|
</PRE>
|
|
<P></P>
|
|
<A NAME="toc43"></A>
|
|
<H3>Slash, objectless sentences</H3>
|
|
<PRE>
|
|
mkSlash : overload {
|
|
</PRE>
|
|
<P></P>
|
|
<P>
|
|
Objectless sentences are used in questions and relative clauses.
|
|
The most common way of constructing them is by using a two-place verb
|
|
with a subject but without an object.
|
|
</P>
|
|
<PRE>
|
|
mkSlash : NP -> V2 -> Slash ; -- 1. (whom) John loves
|
|
</PRE>
|
|
<P></P>
|
|
<P>
|
|
The two-place verb can be separated from the subject by a verb-complement verb.
|
|
</P>
|
|
<PRE>
|
|
mkSlash : NP -> VV -> V2 -> Slash ; -- 2. (whom) John wants to see
|
|
</PRE>
|
|
<P></P>
|
|
<P>
|
|
The missing object can also be the noun phrase in a prepositional phrase.
|
|
</P>
|
|
<PRE>
|
|
mkSlash : Cl -> Prep -> Slash ; -- 3. (with whom) John walks
|
|
</PRE>
|
|
<P></P>
|
|
<P>
|
|
An objectless sentence can be modified by an adverb.
|
|
</P>
|
|
<PRE>
|
|
mkSlash : Slash -> Adv -> Slash -- 4. (whom) John loves today
|
|
} ;
|
|
</PRE>
|
|
<P></P>
|
|
<A NAME="toc44"></A>
|
|
<H2>Lists for coordination</H2>
|
|
<P>
|
|
The rules in this section are very uniform: a list can be built from two or more
|
|
expressions of the same category.
|
|
</P>
|
|
<A NAME="toc45"></A>
|
|
<H3>ListS, sentence lists</H3>
|
|
<PRE>
|
|
mkListS : overload {
|
|
mkListS : S -> S -> ListS ; -- 1. he walks, I run
|
|
mkListS : S -> ListS -> ListS -- 2. John walks, I run, you sleep
|
|
} ;
|
|
</PRE>
|
|
<P></P>
|
|
<A NAME="toc46"></A>
|
|
<H3>ListAdv, adverb lists</H3>
|
|
<PRE>
|
|
mkListAdv : overload {
|
|
mkListAdv : Adv -> Adv -> ListAdv ; -- 1. here, now
|
|
mkListAdv : Adv -> ListAdv -> ListAdv -- 2. to me, here, now
|
|
} ;
|
|
</PRE>
|
|
<P></P>
|
|
<A NAME="toc47"></A>
|
|
<H3>ListAP, adjectival phrase lists</H3>
|
|
<PRE>
|
|
mkListAP : overload {
|
|
mkListAP : AP -> AP -> ListAP ; -- 1. old, big
|
|
mkListAP : AP -> ListAP -> ListAP -- 2. old, big, warm
|
|
} ;
|
|
</PRE>
|
|
<P></P>
|
|
<A NAME="toc48"></A>
|
|
<H3>ListNP, noun phrase lists</H3>
|
|
<PRE>
|
|
mkListNP : overload {
|
|
mkListNP : NP -> NP -> ListNP ; -- 1. John, I
|
|
mkListNP : NP -> ListNP -> ListNP -- 2. John, I, that
|
|
} ;
|
|
</PRE>
|
|
<P></P>
|
|
|
|
<!-- html code generated by txt2tags 2.3 (http://txt2tags.sf.net) -->
|
|
<!-- cmdline: txt2tags -thtml -\-toc api/Constructors.txt -->
|
|
</BODY></HTML>
|