Practical Grammar 8: Difference between revisions

From English Grammar
Jump to navigation Jump to search
← Older edit
No edit summary
 
(23 intermediate revisions by the same user not shown)
Line 2: Line 2:
<font size="3">
<font size="3">


So far, the arguments of verbs have all been determiner phrases (= DPs) like names, pronouns, or Det-N (''the cat'') configurations. But verbs can also take complements of other parts of speech. This week, we will encounter two new cases: namely, complements which are a. prepositional phrases and b. subordinate clauses.
== Complement Clauses ==


== Prepositional Phrases ==
Next, we come to the exciting topic of complement (= subordinate) clauses. Here are two examples:


We want our grammar to generate sentences like the following:
(1) Fred thought that Lilly disappeared<br>
(2) Fred asked whether Lilly disappeared


(1) The cat sat under the table.
There is nothing really special about these structures. As with prepositional phrases, we need
(2) Robin put food on the table.


To achieve this, we need
# lexical items for the new verbs
# lexical items for the new verbs
# a phrase structure rule that creates PPs from a P and a DP
# lexical items for the two complementizers ''that'' and ''whether''
# two new phrase structure rules for verb phrases.
# two new phrase structure rules.


=== The Syntax of PPs ===
We make the following assumptions:
 
We make the following assumptions about the internal and external syntax of PPs:
 
A. The internal structure of the PP
 
a. the P is the head of the PP<br>
b. the DP daughter of the PP bears the grammatical function OBJ
 
B. The external relationships of the PP
 
a. '''c-structure''' the whole PP is treated as just another daughter of the verb phrase whose head is the verb selecting the PP. Thus, in (1), the VP has two daughters, a V and a PP. In (2), the VP has three daughters, a V, a DP, and a PP.<br>
b. '''Grammatical function''' we will only deal with PPs that express locations, for example ''under the table'' and ''on the table''. Accordingly, the verbs taking the PP as argument assign it the grammatical function '''LOC'''.
 
=== The Semantics of PPs expressing locations ===
 
We make the following assumptions about the meaning of locational PPs like ''under the table'':
 
a. the meaning of the DP ''the table'' is a LANDMARK (= orientation point)
b. different prepositions take the landmark as a basis and make different locations, depending on the meaning of the preposition. Thus, ''under'' uses the landmark '''the table''' to create the location '''under the table''', whereas ''on'' makes the location '''on the table''' from the same landmark.
c. Sentence (1) then says that the cat is sitting in the location under the table and sentence (2) says that Robin put the food into the location '''on the table'''.
d. In accordance with this, the whole PP bears the thematic role LOCATION to the verb, and
e. the DP object of the preposition bears the thematic role of LANDMARK to the preposition.
 
 
 
 
 
 
 
(2) translates into xlfg as follows:
 
3. VP → V PP
2. {
3. ↑=↓1;
4. (↑OBL) = ↓2;
5. }
 
Add the following rule for PPs:
 
6.  PP → P DP
7.  {
8.  ↑=↓1;
9.  (↑OBJ) =↓2;
10. }
 
<span style="color: blue>Exercise 7.1</span>
 
1. Open Grammar-7.1<br>
2. Add the two rules above to your grammar.<br>
3. Also add a correctly annotated rule VP -> V DP PP.<br>
2. Make sure that in all your VP rules that have a PP on the right the annotation on the PP is: (↑OBL) = ↓n.<br>
3. Add the prepositions ''to'' on ''on'' to your lexicon as follows:
to P
[PRED: 'to<OBJ>',
  PCASE: OBL_REC]
 
on P
[PRED: 'on<OBJ>',
  PCASE: OBL_LOC]
 
4. Add the following test items:
a. Oscar donated clothes to charity
b. Robin put food on the table
5. Parse.
<br>
 
<span style="color: blue>Exercise 7.2</span>
 
1. Add the following test items:
a. *Oscar donated clothes on charity
b. *Robin put food to the table
2. Parse!<br>
3. The grammar does not return the intended result. What is the problem?<br>
4. The problem is easy to fix with the same technique that we used to enforce subject-verb agreement.<br>
5. Do so and parse!
 
== Complement Clauses ==
 
<span style="color: blue>Exercise 8.1 (based on section 5.1 of the textbook)</span>
 
* Go to <span class="newwin">[https://147.210.117.56 https://147.210.117.56]</span>.
<!-- Grammar5-Ex 7.3-solution. -->
 
* Add the words in the following sentences:
 
(1) Fred thinks that Lilly disappeared<br>
(2) Fred enquires whether Lilly disappeared
 
Implement the following:


# ''that'' and ''whether'' belong to the part of speech C (= complementizer).  
# ''that'' and ''whether'' belong to the part of speech C (= complementizer).  
# Complementizers head CPs.
# A complementizer combines with a following S to form another S.
# CPs take two daughers: a C and an IP. The two daughters are co-heads of the CP.
# The C and the lower S are co-heads of the upper S.
# You need to add a new VP rule which allows a VP to consist of a V and a CP. The CP bears the GF '''COMP'''.
# You need to add a new VP rule which allows a VP to consist of a V and an S. The S bears the GF '''COMP'''.
 
# Complementizers have no PRED value.
* Make any further changes that are necessary to obtain the f-structure (10) on p. 101 for (1), with one exception: your grammar will not contain the information [CLTYPE DECL] in the main f-structure! For sentence (2), you should obtain an f-structure which is identical to that of sentence (1), with the exception that the CLTYPE of the COMP in (2) should be INTER.
 
Check what your grammar predicts for the following examples:
 
(3) *Fred thinks whether Lilly disappeared<br>
(4) *Fred enquires that Lilly disappeared
 
If your grammar licenses these examples, then add information to it that makes it impossible to derive a well formed f-structure for them.
 
Hint: think about the function of each of the complementizers!
 
== Subject Clauses ==
 
''That''- and ''whether''-clauses can also act as subjects:
 
1. That Lilly disappeared sucks<br>
2. Whether Lilly disappeared is unknown
 
Of course, the complementizers are still bound to their clause types:
 
3. *Whether Lilly disappeared sucks<br>
 
<span style="color: blue>Exercise 8.2</span>
 
* Go to <span class="newwin">[https://xlfg.labri.fr/ https://xlfg.labri.fr/]</span>.
<!-- * Open your latest grammar or start with a copy of Grammar5-.Ex8.1-solution. -->
* Add the sentences above to your test items.
* Change the grammar so that it makes the correct predictions for all test items.
 
Your grammar should yield the following Argument Structure for sentence (1):<br>
 
[[File:Sucks.JPG|500px]]
 
Your grammar should yield the following Argument Structure for sentence (2):<br>
 
[[File:Unknown.JPG|500px]]
 
Notes:
 
a. treat ''is'' as a verb which combines with a COMP and an OBJ.<br>
b. treat ''unknown'' as an adjective.
 
 
 
<!--
== The Difference between Defining Equations and Constraining Equations ==
 
There are several types of equations that can be used in annotations. So far, we have encountered the following two:
 
(1) ↑=↓1; <br>
(2) (↑ OBJ) =↓2;
 
These equations are both defining equations.
 
'''Defining equations''' add their information to an f-structure.
 
There is a second kind of equation, which we have not seen yet, but which you will need for the following exercise. These are called constraininig equations.
 
'''Constraininig equations''' test whether their information is contained in an f-structure. They do NOT add the information themselves.
 
Illustration:
 
Case 1:
 
Imagine you have the following defining equation:
 
(↑ TENSE) = pres;
 
* it turns the f-structure [] into the f-structure [TENSE pres], i.e. it adds its information to the f-structure.
* it turns the f-structure [TENSE pres] into the f-structure [TENSE pres], i.e. it adds its information to the f-structure. If the information was already there, the f-structure remains the same.


Case 2:
<span style="color: blue>Exercise 9 </span>


Now, imagine you have the following constraining equation:
1. Open Grammar Grammar 8<br>
2. Implement the analysis for sentences (1)-(2) as described above.<br>
3. Parse. Your output should look exactly like the output decribed in the document Exercise-9-expected-output.pdf on Olat.


(↑ TENSE) =<sub>c</sub> pres
Note that the complementizers in (1) and (2) cannot be exchanged:


* it marks the f-structure [] as ill-formed, since it does not contain the information TENSE pres, i.e. the constraining equation is a test on an f-structure.
(3) *Fred asked that Lilly disappeared<br>
* it marks the f-structure [TENSE pres] as well-formed, but does not change it.
(4) *Fred thought whether Lilly disappeared


When to use a constraining equation:
The reason is that there is an incompatibility of clause type information in (3) and (4): the verb ''thought'' requires a declarative clause as its COMP, but
the word ''whether'' can only head interrogative clauses. In (4), we find the opposite incompatibility.


'''Constraining equations''' are used when one item depends on some other item's adding a particular piece of information to an f-structure.
<span style="color: blue>Exercise 10 </span>


=== The Syntax of Constraining equations in xlfg ===
1. Add the feature CLAUSE_TYPE to the lexical entries that need it so that (3)-(4) are not accepted by the grammar for the reasons stated above, but (1)-(2) stay grammatical.<br>
2. Parse. Your output should look exactly like the output decribed in the document Exercise-9-expected-output.docx on Olat.


Since xlfg does not use subscripts, it uses "==" as constraining equations.


Illustration:


'''<span style="color: red>(↑ TENSE) =<sub>c</sub> pres</span>''' becomes '''<span style="color: blue>(↑ TENSE) == pres;</span>''' in xlfg.
-->
<br>
<br>
<br>
<br>
Line 210: Line 49:
<font size="2">
<font size="2">
<div align="center">
<div align="center">
   [[Practical_Grammar |'''Main page''']]  [[Practical_Grammar_2|'''Week 2''']]  [[Practical_Grammar_3|'''Week 3''']]  [[Practical_Grammar_4|'''Week 4''']]  [[Practical_Grammar_5|'''Week 5''']]  [[Practical_Grammar_6|'''Week 6''']]  [[Practical_Grammar_7_new|'''Week 7''']]  '''Week 8'''  [[Practical_Grammar_9|'''Week 9''']] 
   [[Practical_Grammar |'''Main page''']]  [[Practical_Grammar_2|'''Week 2''']]  [[Practical_Grammar_3|'''Week 3''']]  [[Practical_Grammar_4|'''Week 4''']]  [[Practical_Grammar_5|'''Week 5''']]  [[Practical_Grammar_6|'''Week 6''']]  [[Practical_Grammar_7|'''Week 7''']]  [[Practical_Grammar_8|'''Week 8''']] '''Week 9'''    
</div>
</div>


<!--  [[Practical_Grammar_10|'''Week 10''']]   [[Practical_Grammar_11|'''Week 11''']] [[Practical_Grammar_12|'''Term Paper Project''']] -->
<!--  [[Practical_Grammar_10|'''Week 10''']] [[Practical_Grammar_11|'''Week 11''']] [[Practical_Grammar_12|'''Term Paper Project''']] -->

Latest revision as of 05:28, 16 June 2026

Complement Clauses

Next, we come to the exciting topic of complement (= subordinate) clauses. Here are two examples:

(1) Fred thought that Lilly disappeared
(2) Fred asked whether Lilly disappeared

There is nothing really special about these structures. As with prepositional phrases, we need

  1. lexical items for the new verbs
  2. lexical items for the two complementizers that and whether
  3. two new phrase structure rules.

We make the following assumptions:

  1. that and whether belong to the part of speech C (= complementizer).
  2. A complementizer combines with a following S to form another S.
  3. The C and the lower S are co-heads of the upper S.
  4. You need to add a new VP rule which allows a VP to consist of a V and an S. The S bears the GF COMP.
  5. Complementizers have no PRED value.

Exercise 9

1. Open Grammar Grammar 8
2. Implement the analysis for sentences (1)-(2) as described above.
3. Parse. Your output should look exactly like the output decribed in the document Exercise-9-expected-output.pdf on Olat.

Note that the complementizers in (1) and (2) cannot be exchanged:

(3) *Fred asked that Lilly disappeared
(4) *Fred thought whether Lilly disappeared

The reason is that there is an incompatibility of clause type information in (3) and (4): the verb thought requires a declarative clause as its COMP, but the word whether can only head interrogative clauses. In (4), we find the opposite incompatibility.

Exercise 10

1. Add the feature CLAUSE_TYPE to the lexical entries that need it so that (3)-(4) are not accepted by the grammar for the reasons stated above, but (1)-(2) stay grammatical.
2. Parse. Your output should look exactly like the output decribed in the document Exercise-9-expected-output.docx on Olat.





Retrieved from "https://www.english-linguistics.de/grammarparadise/wiki/index.php?title=Practical_Grammar_8&oldid=5841"