Grammar Writing: Term Project

2018-02-09T11:32:38Z

Nikoschenk: /* Exercise (due March, 11th) */

2018-02-09T11:07:04Z

Nikoschenk: Created blank page

Grammar Writing: Week 11

2018-02-09T11:06:47Z

Nikoschenk:

= Auxiliaries, Identity and the Head-Complement Rule =

So far, all our sentences just have a single finite main verb in them. But, of course English also has sentences with one or more auxiliary verbs and a main verb:

1. Lilly is dancing 
2. Lilly has danced 
3. Lilly will dance 
4. Lilly will be dancing 
5. Lilly will have danced 
6. Lilly will have been dancing 

== Exercise 1 ==

While the sentences above are grammatical, the ones below are ungrammatical:

7. *Lilly is dance 
8. *Lilly has dancing 
9. *Lilly will danced

Fill in the blanks below:

a. The forms of the progressive auxiliary ''be'' select a VP-complement whose VFORM value is: _____ . 
b. The forms of the progressive auxiliary ''have'' select a VP-complement whose VFORM value is: _____ . 
c. The forms of the progressive auxiliary ''will'' select a VP-complement whose VFORM value is: _____ .

== Exercise 2 ==

Add lexical entries to your grammar which license all the sentences (1)-(6), but disallow (7)-(9).

Make the following assumptions:

# All auxiliaries belong to part of speech ''verb''.
# Some auxiliaries have different inflectional forms (e.g. ''be, is, been'').
# The modal auxiliaries (e.g. ''must, can'') and ''will'' only have finite forms.
# If an auxiliary has more than one form, then you need a lexical entry for each different form!
# Every auxiliary selects a VP-complement with a particular VFORM value (see above).

NOTE: in order to complete this excercise, you do NOT need to add any syntactic rules. All you need are the lexical entries of the auxiliaries!

CAUTION: make sure that your grammar does NOT license any of the following sentences:

8.* Lilly is liking. 
9. *Lilly has liked. 
10. *Lilly will like. 

Hint: the auxiliaries should make sure that their complement is itself COMPS-complete.

= Specifiers =

Most of the NPs we have modelled so far were either names or pronouns. But like VPs, APs, PPs, and Ss, NPs can of course also consist of more than one word:

11. the cat 
12. a cat

These NPs all consist of a '''determiner''' D and a noun. Our grammar up to this point does not allow us to license such multi-word NPs. But, it is easy to do so. For that purpose, we will postulate a new grammatical function, namely the SPECIFIER. We will assume that the words ''cat'' and ''cats'' in (11)-(12) are unspecific as to how many (of the) cats the speaker wishes to identify (one vs. more than one) and whether a definite cat is referred about or not (''the'' vs ''a''). The function of the determiner is to specify this kind of information. For this reason, we assume that the NPs in (11)-(12) have the form of SPECIFIER + HEAD.

== Exercise 3 (due date: Sunday, 8 p.m. on January 14th 2018) ==

# Add the part of speech ''det'' to the same place in the type hierarchy where the other parts of speech are already defined.
# Add the feature SPR with value ''list'' to the same place in the type hierarchy where the feature COMPS is already defined.
# Add lexical entries for the determiners ''the, a,'' and ''those'' to the lexicon. Determines do not need any complements and any specifiers.
# Delete the old lexical entry for the word ''cat''.
# Add a new lexical entry for ''cat'' as follows: there word takes no complements but does take one D as specifier.

Upload your grammar to make sure that the structure of the determiners are compatible with what the single member of the SPR list of the word ''cat''.

== Exercise 4 (due date: Sunday, 8 p.m. on January 14th 2018) ==

If everything went well, then at this point your grammar contains the noun ''cat'' and the three determiners ''the, a,'' and ''those'' in such a form that the determiners make possible specifiers of ''cat''. All that is missing for the grammar to license NPs of the form SPR + HEAD is a new phrase structure for NPs. Like the VP-rule, the NP-rule will combine a HEAD daughter with a non-head daughter, but there are 2 differences: a. this time the second daughter will be a specifier rather than a complement; (b) the order of the two daughters is reversed: the first daughter is the specifier and the second one the head daughter!

With these preparations, write a new phrase structure rule, the head_specifier_rule, as follows:

'''Note that you should download a new version of the tool (it contains a bug fix for the rule we have dealt with last time in class, you can simply download and replace the .jar file)'''
direct link to new version: [https://github.com/Nikoschenk/desktoptrale/blob/master/desktoptrale/desktoptrale.jar]

# The mother is a phrase with empty COMPS and SPR lists.
# The second daughter is a noun word with empty COMPS and <D> as a SPR.
# The first daughter is a determiner word.
# The HEAD value of the mother is identical to that of the second daughter.

The following three versions should all work:
Implement one of them but note the differences in the elements that share identical properties.

[[File:Hsr1.png]]
[[File:Hsr2.png]]
[[File:Hsr3.png]]

= Modifiers =

Our grammar already can handle a lot of features of English syntax:

# Main verbs of different inflectional types
# Subjects and complements
# Auxiliaries
# Prepositional phrases
# Simple noun phrases

One important grammatical function is still missing, however, namely '''modifiers'''. Modifiers (also called '''adjuncts''') differ from arguments in two important respects:

# Arguments (= subjects and complements) are selected by a head and are usually obligatory.
# Adjuncts in turn select the head and thus are always optional.

== Adjectival modifiers ==

The most typical modifiers of nouns are adjectives:

(1) the student 
(2) the '''clever''' student

Note that adjectives cannot modify verbs:

(3) Lilly ran 
(4) * Lilly '''clever''' ran.

Note that the adjectival modifier has to appear between the specifier and the head:

(5) *'''clever''' the student

Also note that only common nouns (i.e. the ones that refer to kinds of things) can be modified, nouns that refer to individuals cannot:

(6) *'''clever''' she

=== Exercise ===

We want to extend our grammar so that it predicts the grammaticality judgments of the expressions (1)-(6) above. To achieve this, we need lexical entries for adjectives and a new grammar rule.

A. Add a lexical entry for the word ''clever'' to the grammar. Give it the empty list of specifiers and complements. 
B. Add a new phrase structure rule ''Adjective-Noun-Phrase'' with the following properties:

# The first daughter is an adjective word.
# The second daughter is a noun sign with empty COMPS.
# The mother is a phrase with empty COMPS.
# The HEAD value of the mother is identical to that of the second daughter.

The way it is written, the rule will overgenerate (i.e. allow expressions that are ungrammatical). For instance, it will wrongly license the expressions (5) and (6) above. However, the problem is easily solved. Think about the grammatical difference between the specifier lists of common nouns on the one hand and those of names and pronouns on the other hand. Then ask yourself the following questions:

# Adjectives can modify common nouns but not names and pronouns: what value of the SPR list of the head daughter in the ''Adjective-Noun-Phrase'' will predict this state of affairs.
# What value should the SPR list of the mother have in order to allow a determiner to be added in front of a ''Adjective-Noun-Phrase''?

=== Exercise ===

What do you have to add to the grammar for it to license the strings below:

(7) the clever young student 
(8) the clever tall young student

How many adjectives can a noun phrase have in principle in this grammar?

== Prepositional Phrase Modifiers ==

Many verbs can be modified by Prepositional Phrases:

(9) Lilly arrived. 
(10) Lilly arrived ''on Monday''.

=== Exercise ===

It is very easy to extend our grammar to license examples like (10). Do the following:

A. Add the word ''on'' to the lexicon. It is a regular preposition (prepositions do not take specifiers). 
B. Add the word ''Monday'' to the lexicon. Treat the word like a proper name. 
C. Add a new rule to the grammar, as follows:

# The first daughter is a verb sign with empty COMPS.
# The second daughter is a PP.
# The mother is a phrase with empty COMPS.
# The HEAD value of the mother is identical to that of the first daughter.

=== Exercise ===

Note that PP modifiers can also be added after a noun:

(11) The party. 
(12) The party ''on Monday''.

Add a rule to the grammar which licenses example (12). The rule will be almost identical to the rule in the previous exercise.

=== Exercise ===

Add the lexical entry for the adjective ''rousing'' to the grammar and nothing else. Does your grammar now license the sentence below:

(13) The rousing party on Monday

If your grammar licenses the sentence, how many trees to you get? Explain your answer!

= Agreement =

In case you are working already on agreement features, please make sure to '''spell out''' numbers in type names, i.e. "non_3s" should be "nonthirdsingular", and so on.

 
 
Navigation:
<div align="center">
[[Grammar_Writing|'''Main Page''']]   [[Grammar_Writing:_Week_2| '''Week 2''']]    [[Grammar_Writing:_Week_3| '''Week 3''']]    [[Grammar_Writing:_Week_4| '''Week 4''']]    [[Grammar_Writing:_Week_5| '''Week 5''']]    [[Grammar_Writing:_Week_6| '''Week 6''']]    [[Grammar_Writing:_Week_7| '''Week 7''']]   [[Grammar_Writing:_Week_8| '''Week 8''']]    [[Grammar_Writing:_Week_9| '''Week 9''']]    [[Grammar_Writing:_Week_10| '''Week 10''']]    '''Week 11'''    [[Grammar_Writing:_Term_Project| '''Term Project''']]

Grammar Writing: Week 11

2018-01-26T22:02:51Z

Nikoschenk:

= Auxiliaries, Identity and the Head-Complement Rule =

So far, all our sentences just have a single finite main verb in them. But, of course English also has sentences with one or more auxiliary verbs and a main verb:

1. Lilly is dancing 
2. Lilly has danced 
3. Lilly will dance 
4. Lilly will be dancing 
5. Lilly will have danced 
6. Lilly will have been dancing 

== Exercise 1 ==

While the sentences above are grammatical, the ones below are ungrammatical:

7. *Lilly is dance 
8. *Lilly has dancing 
9. *Lilly will danced

Fill in the blanks below:

a. The forms of the progressive auxiliary ''be'' select a VP-complement whose VFORM value is: _____ . 
b. The forms of the progressive auxiliary ''have'' select a VP-complement whose VFORM value is: _____ . 
c. The forms of the progressive auxiliary ''will'' select a VP-complement whose VFORM value is: _____ .

== Exercise 2 ==

Add lexical entries to your grammar which license all the sentences (1)-(6), but disallow (7)-(9).

Make the following assumptions:

# All auxiliaries belong to part of speech ''verb''.
# Some auxiliaries have different inflectional forms (e.g. ''be, is, been'').
# The modal auxiliaries (e.g. ''must, can'') and ''will'' only have finite forms.
# If an auxiliary has more than one form, then you need a lexical entry for each different form!
# Every auxiliary selects a VP-complement with a particular VFORM value (see above).

NOTE: in order to complete this excercise, you do NOT need to add any syntactic rules. All you need are the lexical entries of the auxiliaries!

CAUTION: make sure that your grammar does NOT license any of the following sentences:

8.* Lilly is liking. 
9. *Lilly has liked. 
10. *Lilly will like. 

Hint: the auxiliaries should make sure that their complement is itself COMPS-complete.

= Specifiers =

Most of the NPs we have modelled so far were either names or pronouns. But like VPs, APs, PPs, and Ss, NPs can of course also consist of more than one word:

11. the cat 
12. a cat

These NPs all consist of a '''determiner''' D and a noun. Our grammar up to this point does not allow us to license such multi-word NPs. But, it is easy to do so. For that purpose, we will postulate a new grammatical function, namely the SPECIFIER. We will assume that the words ''cat'' and ''cats'' in (11)-(12) are unspecific as to how many (of the) cats the speaker wishes to identify (one vs. more than one) and whether a definite cat is referred about or not (''the'' vs ''a''). The function of the determiner is to specify this kind of information. For this reason, we assume that the NPs in (11)-(12) have the form of SPECIFIER + HEAD.

== Exercise 3 (due date: Sunday, 8 p.m. on January 14th 2018) ==

# Add the part of speech ''det'' to the same place in the type hierarchy where the other parts of speech are already defined.
# Add the feature SPR with value ''list'' to the same place in the type hierarchy where the feature COMPS is already defined.
# Add lexical entries for the determiners ''the, a,'' and ''those'' to the lexicon. Determines do not need any complements and any specifiers.
# Delete the old lexical entry for the word ''cat''.
# Add a new lexical entry for ''cat'' as follows: there word takes no complements but does take one D as specifier.

Upload your grammar to make sure that the structure of the determiners are compatible with what the single member of the SPR list of the word ''cat''.

== Exercise 4 (due date: Sunday, 8 p.m. on January 14th 2018) ==

If everything went well, then at this point your grammar contains the noun ''cat'' and the three determiners ''the, a,'' and ''those'' in such a form that the determiners make possible specifiers of ''cat''. All that is missing for the grammar to license NPs of the form SPR + HEAD is a new phrase structure for NPs. Like the VP-rule, the NP-rule will combine a HEAD daughter with a non-head daughter, but there are 2 differences: a. this time the second daughter will be a specifier rather than a complement; (b) the order of the two daughters is reversed: the first daughter is the specifier and the second one the head daughter!

With these preparations, write a new phrase structure rule, the head_specifier_rule, as follows:

'''Note that you should download a new version of the tool (it contains a bug fix for the rule we have dealt with last time in class, you can simply download and replace the .jar file)'''
direct link to new version: [https://github.com/Nikoschenk/desktoptrale/blob/master/desktoptrale/desktoptrale.jar]

# The mother is a phrase with empty COMPS and SPR lists.
# The second daughter is a noun word with empty COMPS and <D> as a SPR.
# The first daughter is a determiner word.
# The HEAD value of the mother is identical to that of the second daughter.

The following three versions should all work:
Implement one of them but note the differences in the elements that share identical properties.

[[File:Hsr1.png]]
[[File:Hsr2.png]]
[[File:Hsr3.png]]

= Modifiers =

Our grammar already can handle a lot of features of English syntax:

# Main verbs of different inflectional types
# Subjects and complements
# Auxiliaries
# Prepositional phrases
# Simple noun phrases

One important grammatical function is still missing, however, namely '''modifiers'''. Modifiers (also called '''adjuncts''') differ from arguments in two important respects:

# Arguments (= subjects and complements) are selected by a head and are usually obligatory.
# Adjuncts in turn select the head and thus are always optional.

== Adjectival modifiers ==

The most typical modifiers of nouns are adjectives:

(1) the student 
(2) the '''clever''' student

Note that adjectives cannot modify verbs:

(3) Lilly ran 
(4) * Lilly '''clever''' ran.

Note that the adjectival modifier has to appear between the specifier and the head:

(5) *'''clever''' the student

Also note that only common nouns (i.e. the ones that refer to kinds of things) can be modified, nouns that refer to individuals cannot:

(6) *'''clever''' she

=== Exercise ===

We want to extend our grammar so that it predicts the grammaticality judgments of the expressions (1)-(6) above. To achieve this, we need lexical entries for adjectives and a new grammar rule.

A. Add a lexical entry for the word ''clever'' to the grammar. Give it the empty list of specifiers and complements. 
B. Add a new phrase structure rule ''Adjective-Noun-Phrase'' with the following properties:

# The first daughter is an adjective word.
# The second daughter is a noun sign with empty COMPS.
# The mother is a phrase with empty COMPS.
# The HEAD value of the mother is identical to that of the second daughter.

The way it is written, the rule will overgenerate (i.e. allow expressions that are ungrammatical). For instance, it will wrongly license the expressions (5) and (6) above. However, the problem is easily solved. Think about the grammatical difference between the specifier lists of common nouns on the one hand and those of names and pronouns on the other hand. Then ask yourself the following questions:

# Adjectives can modify common nouns but not names and pronouns: what value of the SPR list of the head daughter in the ''Adjective-Noun-Phrase'' will predict this state of affairs.
# What value should the SPR list of the mother have in order to allow a determiner to be added in front of a ''Adjective-Noun-Phrase''?

=== Exercise ===

What do you have to add to the grammar for it to license the strings below:

(7) the clever young student 
(8) the clever tall young student

How many adjectives can a noun phrase have in principle in this grammar?

== Prepositional Phrase Modifiers ==

Many verbs can be modified by Prepositional Phrases:

(9) Lilly arrived. 
(10) Lilly arrived ''on Monday''.

=== Exercise ===

It is very easy to extend our grammar to license examples like (10). Do the following:

A. Add the word ''on'' to the lexicon. It is a regular preposition (prepositions do not take specifiers). 
B. Add the word ''Monday'' to the lexicon. Treat the word like a proper name. 
C. Add a new rule to the grammar, as follows:

# The first daughter is a verb sign with empty COMPS.
# The second daughter is a PP.
# The mother is a phrase with empty COMPS.
# The HEAD value of the mother is identical to that of the first daughter.

=== Exercise ===

Note that PP modifiers can also be added after a noun:

(11) The party. 
(12) The party ''on Monday''.

Add a rule to the grammar which licenses example (12). The rule will be almost identical to the rule in the previous exercise.

=== Exercise ===

Add the lexical entry for the adjective ''rousing'' to the grammar and nothing else. Does your grammar now license the sentence below:

(13) The rousing party on Monday

If your grammar licenses the sentence, how many trees to you get? Explain your answer!

= Agreement =

In case you are working already on agreement features, please make sure to '''spell out''' numbers in type names, i.e. "non_3s" should be "nonthirdsingular", and so on.

 
 
Navigation:
<div align="center">
[[Grammar_Writing|'''Main Page''']]   [[Grammar_Writing:_Week_2| '''Week 2''']]    [[Grammar_Writing:_Week_3| '''Week 3''']]    [[Grammar_Writing:_Week_4| '''Week 4''']]    [[Grammar_Writing:_Week_5| '''Week 5''']]    [[Grammar_Writing:_Week_6| '''Week 6''']]    [[Grammar_Writing:_Week_7| '''Week 7''']]   [[Grammar_Writing:_Week_8| '''Week 8''']]    [[Grammar_Writing:_Week_9| '''Week 9''']]    [[Grammar_Writing:_Week_10| '''Week 10''']]    '''Week 11'''

Grammar Writing: Week 11

2018-01-17T15:18:29Z

Nikoschenk: /* Exercise 4 (due date: Sunday, 8 p.m. on January 14th 2018) */

= Auxiliaries, Identity and the Head-Complement Rule =

So far, all our sentences just have a single finite main verb in them. But, of course English also has sentences with one or more auxiliary verbs and a main verb:

1. Lilly is dancing 
2. Lilly has danced 
3. Lilly will dance 
4. Lilly will be dancing 
5. Lilly will have danced 
6. Lilly will have been dancing 

== Exercise 1 ==

While the sentences above are grammatical, the ones below are ungrammatical:

7. *Lilly is dance 
8. *Lilly has dancing 
9. *Lilly will danced

Fill in the blanks below:

a. The forms of the progressive auxiliary ''be'' select a VP-complement whose VFORM value is: _____ . 
b. The forms of the progressive auxiliary ''have'' select a VP-complement whose VFORM value is: _____ . 
c. The forms of the progressive auxiliary ''will'' select a VP-complement whose VFORM value is: _____ .

== Exercise 2 ==

Add lexical entries to your grammar which license all the sentences (1)-(6), but disallow (7)-(9).

Make the following assumptions:

# All auxiliaries belong to part of speech ''verb''.
# Some auxiliaries have different inflectional forms (e.g. ''be, is, been'').
# The modal auxiliaries (e.g. ''must, can'') and ''will'' only have finite forms.
# If an auxiliary has more than one form, then you need a lexical entry for each different form!
# Every auxiliary selects a VP-complement with a particular VFORM value (see above).

NOTE: in order to complete this excercise, you do NOT need to add any syntactic rules. All you need are the lexical entries of the auxiliaries!

CAUTION: make sure that your grammar does NOT license any of the following sentences:

8.* Lilly is liking. 
9. *Lilly has liked. 
10. *Lilly will like. 

Hint: the auxiliaries should make sure that their complement is itself COMPS-complete.

= Specifiers =

Most of the NPs we have modelled so far were either names or pronouns. But like VPs, APs, PPs, and Ss, NPs can of course also consist of more than one word:

11. the cat 
12. a cat

These NPs all consist of a '''determiner''' D and a noun. Our grammar up to this point does not allow us to license such multi-word NPs. But, it is easy to do so. For that purpose, we will postulate a new grammatical function, namely the SPECIFIER. We will assume that the words ''cat'' and ''cats'' in (11)-(12) are unspecific as to how many (of the) cats the speaker wishes to identify (one vs. more than one) and whether a definite cat is referred about or not (''the'' vs ''a''). The function of the determiner is to specify this kind of information. For this reason, we assume that the NPs in (11)-(12) have the form of SPECIFIER + HEAD.

== Exercise 3 (due date: Sunday, 8 p.m. on January 14th 2018) ==

# Add the part of speech ''det'' to the same place in the type hierarchy where the other parts of speech are already defined.
# Add the feature SPR with value ''list'' to the same place in the type hierarchy where the feature COMPS is already defined.
# Add lexical entries for the determiners ''the, a,'' and ''those'' to the lexicon. Determines do not need any complements and any specifiers.
# Delete the old lexical entry for the word ''cat''.
# Add a new lexical entry for ''cat'' as follows: there word takes no complements but does take one D as specifier.

Upload your grammar to make sure that the structure of the determiners are compatible with what the single member of the SPR list of the word ''cat''.

== Exercise 4 (due date: Sunday, 8 p.m. on January 14th 2018) ==

If everything went well, then at this point your grammar contains the noun ''cat'' and the three determiners ''the, a,'' and ''those'' in such a form that the determiners make possible specifiers of ''cat''. All that is missing for the grammar to license NPs of the form SPR + HEAD is a new phrase structure for NPs. Like the VP-rule, the NP-rule will combine a HEAD daughter with a non-head daughter, but there are 2 differences: a. this time the second daughter will be a specifier rather than a complement; (b) the order of the two daughters is reversed: the first daughter is the specifier and the second one the head daughter!

With these preparations, write a new phrase structure rule, the head_specifier_rule, as follows:

'''Note that you should download a new version of the tool (it contains a bug fix for the rule we have dealt with last time in class, you can simply download and replace the .jar file)'''
direct link to new version: [https://github.com/Nikoschenk/desktoptrale/blob/master/desktoptrale/desktoptrale.jar]

# The mother is a phrase with empty COMPS and SPR lists.
# The second daughter is a noun word with empty COMPS and <D> as a SPR.
# The first daughter is a determiner word.
# The HEAD value of the mother is identical to that of the second daughter.

The following three versions should all work:
Implement one of them but note the differences in the elements that share identical properties.

[[File:Hsr1.png]]
[[File:Hsr2.png]]
[[File:Hsr3.png]]

= Modifiers =

Our grammar already can handle a lot of features of English syntax:

# Main verbs of different inflectional types
# Subjects and complements
# Auxiliaries
# Prepositional phrases
# Simple noun phrases

One important grammatical function is still missing, however, namely '''modifiers'''. Modifiers (also called '''adjuncts''') differ from arguments in two important respects:

# Arguments (= subjects and complements) are selected by a head and are usually obligatory.
# Adjuncts in turn select the head and thus are always optional.

== Adjectival modifiers ==

The most typical modifiers of nouns are adjectives:

(1) the student 
(2) the '''clever''' student

Note that adjectives cannot modify verbs:

(3) Lilly ran 
(4) * Lilly '''clever''' ran.

Note that the adjectival modifier has to appear between the specifier and the head:

(5) *'''clever''' the student

Also note that only common nouns (i.e. the ones that refer to kinds of things) can be modified, nouns that refer to individuals cannot:

(6) *'''clever''' she

=== Exercise ===

We want to extend our grammar so that it predicts the grammaticality judgments of the expressions (1)-(6) above. To achieve this, we need lexical entries for adjectives and a new grammar rule.

A. Add a lexical entry for the word ''clever'' to the grammar. Give it the empty list of specifiers and complements. 
B. Add a new phrase structure rule ''Adjective-Noun-Phrase'' with the following properties:

# The first daughter is an adjective word.
# The second daughter is a noun sign with empty COMPS.
# The mother is a phrase with empty COMPS.
# The HEAD value of the mother is identical to that of the second daughter.

The way it is written, the rule will overgenerate (i.e. allow expressions that are ungrammatical). For instance, it will wrongly license the expressions (5) and (6) above. However, the problem is easily solved. Think about the grammatical difference between the specifier lists of common nouns on the one hand and those of names and pronouns on the other hand. Then ask yourself the following questions:

# Adjectives can modify common nouns but not names and pronouns: what value of the SPR list of the head daughter in the ''Adjective-Noun-Phrase'' will predict this state of affairs.
# What value should the SPR list of the mother have in order to allow a determiner to be added in front of a ''Adjective-Noun-Phrase''?

== Prepositional Phrase Modifiers ==

Many verbs can be modified by Prepositional Phrases:

(7) Lilly arrived. 
(8) Lilly arrived ''on Monday''.

== Exercise ==

It is very easy to extend our grammar to license examples like (8). Do the following:

A. Add the word ''on'' to the lexicon. It is a regular preposition (prepositions do not take specifiers). 
B. Add the word ''Monday'' to the lexicon. Treat the word like a proper name. 
C. Add a new rule to the grammar, as follows:

# The first daughter is a verb sign with empty COMPS.
# The second daughter is a PP.
# The mother is a phrase with empty COMPS.
# The HEAD value of the mother is identical to that of the first daughter.

== Exercise ==

Note that PP modifiers can also be added after a noun:

(9) The party. 
(10) The party ''on Monday''.

Add a rule to the grammar which licenses example (10). The rule will be almost identical to the rule in the previous exercise.

 
 
Navigation:
<div align="center">
[[Grammar_Writing|'''Main Page''']]   [[Grammar_Writing:_Week_2| '''Week 2''']]    [[Grammar_Writing:_Week_3| '''Week 3''']]    [[Grammar_Writing:_Week_4| '''Week 4''']]    [[Grammar_Writing:_Week_5| '''Week 5''']]    [[Grammar_Writing:_Week_6| '''Week 6''']]    [[Grammar_Writing:_Week_7| '''Week 7''']]   [[Grammar_Writing:_Week_8| '''Week 8''']]    [[Grammar_Writing:_Week_9| '''Week 9''']]    [[Grammar_Writing:_Week_10| '''Week 10''']]    '''Week 11'''

Grammar Writing: Week 11

2018-01-17T15:17:32Z

Nikoschenk: /* Exercise 4 (due date: Sunday, 8 p.m. on January 14th 2018) */

= Auxiliaries, Identity and the Head-Complement Rule =

So far, all our sentences just have a single finite main verb in them. But, of course English also has sentences with one or more auxiliary verbs and a main verb:

1. Lilly is dancing 
2. Lilly has danced 
3. Lilly will dance 
4. Lilly will be dancing 
5. Lilly will have danced 
6. Lilly will have been dancing 

== Exercise 1 ==

While the sentences above are grammatical, the ones below are ungrammatical:

7. *Lilly is dance 
8. *Lilly has dancing 
9. *Lilly will danced

Fill in the blanks below:

a. The forms of the progressive auxiliary ''be'' select a VP-complement whose VFORM value is: _____ . 
b. The forms of the progressive auxiliary ''have'' select a VP-complement whose VFORM value is: _____ . 
c. The forms of the progressive auxiliary ''will'' select a VP-complement whose VFORM value is: _____ .

== Exercise 2 ==

Add lexical entries to your grammar which license all the sentences (1)-(6), but disallow (7)-(9).

Make the following assumptions:

# All auxiliaries belong to part of speech ''verb''.
# Some auxiliaries have different inflectional forms (e.g. ''be, is, been'').
# The modal auxiliaries (e.g. ''must, can'') and ''will'' only have finite forms.
# If an auxiliary has more than one form, then you need a lexical entry for each different form!
# Every auxiliary selects a VP-complement with a particular VFORM value (see above).

NOTE: in order to complete this excercise, you do NOT need to add any syntactic rules. All you need are the lexical entries of the auxiliaries!

CAUTION: make sure that your grammar does NOT license any of the following sentences:

8.* Lilly is liking. 
9. *Lilly has liked. 
10. *Lilly will like. 

Hint: the auxiliaries should make sure that their complement is itself COMPS-complete.

= Specifiers =

Most of the NPs we have modelled so far were either names or pronouns. But like VPs, APs, PPs, and Ss, NPs can of course also consist of more than one word:

11. the cat 
12. a cat

These NPs all consist of a '''determiner''' D and a noun. Our grammar up to this point does not allow us to license such multi-word NPs. But, it is easy to do so. For that purpose, we will postulate a new grammatical function, namely the SPECIFIER. We will assume that the words ''cat'' and ''cats'' in (11)-(12) are unspecific as to how many (of the) cats the speaker wishes to identify (one vs. more than one) and whether a definite cat is referred about or not (''the'' vs ''a''). The function of the determiner is to specify this kind of information. For this reason, we assume that the NPs in (11)-(12) have the form of SPECIFIER + HEAD.

== Exercise 3 (due date: Sunday, 8 p.m. on January 14th 2018) ==

# Add the part of speech ''det'' to the same place in the type hierarchy where the other parts of speech are already defined.
# Add the feature SPR with value ''list'' to the same place in the type hierarchy where the feature COMPS is already defined.
# Add lexical entries for the determiners ''the, a,'' and ''those'' to the lexicon. Determines do not need any complements and any specifiers.
# Delete the old lexical entry for the word ''cat''.
# Add a new lexical entry for ''cat'' as follows: there word takes no complements but does take one D as specifier.

Upload your grammar to make sure that the structure of the determiners are compatible with what the single member of the SPR list of the word ''cat''.

== Exercise 4 (due date: Sunday, 8 p.m. on January 14th 2018) ==

If everything went well, then at this point your grammar contains the noun ''cat'' and the three determiners ''the, a,'' and ''those'' in such a form that the determiners make possible specifiers of ''cat''. All that is missing for the grammar to license NPs of the form SPR + HEAD is a new phrase structure for NPs. Like the VP-rule, the NP-rule will combine a HEAD daughter with a non-head daughter, but there are 2 differences: a. this time the second daughter will be a specifier rather than a complement; (b) the order of the two daughters is reversed: the first daughter is the specifier and the second one the head daughter!

With these preparations, write a new phrase structure rule, the head_specifier_rule, as follows:

'''Note that you should download a new version of the tool (it contains a bug fix for the rule we have dealt with last time in class, you can simply download and replace the .jar file)'''
direct link to new version: [https://github.com/Nikoschenk/desktoptrale/blob/master/desktoptrale/desktoptrale.jar]

# The mother is a phrase with empty COMPS and SPR lists.
# The second daughter is a noun word with empty COMPS and <D> as a SPR.
# The first daughter is a determiner word.
# The HEAD value of the mother is identical to that of the second daughter.

The following three versions should all work:

[[File:Hsr1.png]]
[[File:Hsr2.png]]
[[File:Hsr3.png]]

= Modifiers =

Our grammar already can handle a lot of features of English syntax:

# Main verbs of different inflectional types
# Subjects and complements
# Auxiliaries
# Prepositional phrases
# Simple noun phrases

One important grammatical function is still missing, however, namely '''modifiers'''. Modifiers (also called '''adjuncts''') differ from arguments in two important respects:

# Arguments (= subjects and complements) are selected by a head and are usually obligatory.
# Adjuncts in turn select the head and thus are always optional.

== Adjectival modifiers ==

The most typical modifiers of nouns are adjectives:

(1) the student 
(2) the '''clever''' student

Note that adjectives cannot modify verbs:

(3) Lilly ran 
(4) * Lilly '''clever''' ran.

Note that the adjectival modifier has to appear between the specifier and the head:

(5) *'''clever''' the student

Also note that only common nouns (i.e. the ones that refer to kinds of things) can be modified, nouns that refer to individuals cannot:

(6) *'''clever''' she

=== Exercise ===

We want to extend our grammar so that it predicts the grammaticality judgments of the expressions (1)-(6) above. To achieve this, we need lexical entries for adjectives and a new grammar rule.

A. Add a lexical entry for the word ''clever'' to the grammar. Give it the empty list of specifiers and complements. 
B. Add a new phrase structure rule ''Adjective-Noun-Phrase'' with the following properties:

# The first daughter is an adjective word.
# The second daughter is a noun sign with empty COMPS.
# The mother is a phrase with empty COMPS.
# The HEAD value of the mother is identical to that of the second daughter.

The way it is written, the rule will overgenerate (i.e. allow expressions that are ungrammatical). For instance, it will wrongly license the expressions (5) and (6) above. However, the problem is easily solved. Think about the grammatical difference between the specifier lists of common nouns on the one hand and those of names and pronouns on the other hand. Then ask yourself the following questions:

# Adjectives can modify common nouns but not names and pronouns: what value of the SPR list of the head daughter in the ''Adjective-Noun-Phrase'' will predict this state of affairs.
# What value should the SPR list of the mother have in order to allow a determiner to be added in front of a ''Adjective-Noun-Phrase''?

== Prepositional Phrase Modifiers ==

Many verbs can be modified by Prepositional Phrases:

(7) Lilly arrived. 
(8) Lilly arrived ''on Monday''.

== Exercise ==

It is very easy to extend our grammar to license examples like (8). Do the following:

A. Add the word ''on'' to the lexicon. It is a regular preposition (prepositions do not take specifiers). 
B. Add the word ''Monday'' to the lexicon. Treat the word like a proper name. 
C. Add a new rule to the grammar, as follows:

# The first daughter is a verb sign with empty COMPS.
# The second daughter is a PP.
# The mother is a phrase with empty COMPS.
# The HEAD value of the mother is identical to that of the first daughter.

== Exercise ==

Note that PP modifiers can also be added after a noun:

(9) The party. 
(10) The party ''on Monday''.

Add a rule to the grammar which licenses example (10). The rule will be almost identical to the rule in the previous exercise.

 
 
Navigation:
<div align="center">
[[Grammar_Writing|'''Main Page''']]   [[Grammar_Writing:_Week_2| '''Week 2''']]    [[Grammar_Writing:_Week_3| '''Week 3''']]    [[Grammar_Writing:_Week_4| '''Week 4''']]    [[Grammar_Writing:_Week_5| '''Week 5''']]    [[Grammar_Writing:_Week_6| '''Week 6''']]    [[Grammar_Writing:_Week_7| '''Week 7''']]   [[Grammar_Writing:_Week_8| '''Week 8''']]    [[Grammar_Writing:_Week_9| '''Week 9''']]    [[Grammar_Writing:_Week_10| '''Week 10''']]    '''Week 11'''

File:Hsr3.png

2018-01-17T15:16:16Z

Nikoschenk:

File:Hsr2.png

2018-01-17T15:16:06Z

Nikoschenk:

Nikoschenk:

= More Complement Types =

So far, we have only encountered three kinds of verbs, intransitive, transitive, and ditransitive verbs:

# dances: COMPS <>
# likes: COMPS <NP>
# gives: COMPS <NP,NP>

But, there are many more kinds of verbs in English. Restricting ourselves to verbs with one complement, we find the following:

# likes: COMPS <NP>
# is: COMPS <AP>
# depends: COMPS <PP>
# believes: COMPS <sentence>

Task 1:
Make sure that your type hierarchy contains the parts of speech: noun, verb, adj, and prep. If it lacks one or more of these parts of speech, then add them.

Task 2:
Add the following words to the lexicon, giving each an empty COMPS list: she, her, happy.

Task 3:
Add the preposition ''on'' to the lexicon with COMPS <NP>.

Task 4:
If necessary, add lexical entries for the verbs likes, is, depends, and believes. Make sure that each verb has the COMPS list given above!

Question
We want to write rules that license all the VPs below. How many rules do we need for each VP and how many VP rules do we need altogether?

# likes her. Number of VP rules: __
# is happy. Number of VP rules: __
# depends on her. Number of VP rules: __
# believes she is happy. Number of VP rules: __

Total number of VP rules needed: __

Task 5
Write the VP rules needed to license the VPs above. For the VP ''depends on her'', you will need to write a PP rule as well.

= The greatest thing since the invention of sliced bread: the concept '''is the same as''' =

Clearly, it is very tedious to write a different VP rule for every verb type that differs from the other verb types from the kind of complement that verb needs to combine with. But actually, things are much worse than this:

# Not only do verbs with one complement differ from each other in what kind of complement they want to combine with, but the verbs that take two complements do, too. When you count all the verb types, you end up with several dozens, which means that you will also need several dozens different VP rules!
# In addition, nouns, prepositions, and adjectives also can take complements and may differ from each other in the complements they demand. For these, you will need many different rules for NPs, PPs, and APs in addition to the VP rules.
# Altogether, you end up with a grammar of more than a hundred rules for phrases.
# SCREAM: THERE HAS TO BE A BETTER WAY TO HANDLE THIS PROBLEM!

= Grammar 4 (Moving the COMPS Feature from Sign to Syntax) =

<nowiki>
type_hierarchy
bot
sign syntax:syntax
phrase dtrs:list
word
list
ne_list tl:list hd:bot
e_list
vform
fin
nonfin
prp
pfp
base
pos
verb vform:vform
noun case:case
s
prep
syntax head:pos comps:list
case
acc
nom
.
</nowiki>

= Homework, due Monday, Dec 18th =

''' Download the new desktoptrale version which contains a bug fix for the save issue encountered in class''' from this URL
[https://github.com/Nikoschenk/desktoptrale/tree/master/desktoptrale] (and delete the old one).

HOMEWORK
Load the previous type hierarchy and make the necessary modifications to your grammar (to the types, the lexicon and to the rule set) so as to recognize the following sentences:

# Donald likes her.
# Donald is funny.
# Donald depends on her.
# Donald believes the earth is flat.