Syntax lectures period 4
Lecture week 1: Agreement
Chapter 5 is a chapter on agreement. A bigger issue here is language change.
Why is there agreement?
No one knows why agreement occurs. Agreement is meaningless, the -s on the verb has no meaning.
‘I say to him’ means the same as ‘I says to him’.
Another point is that in Italian, you can leave out the subject. In English, this is not possible. In Italian,
the -s can help you find the missing subject. This is not the case in English, the -s really has no
purpose. Maybe agreement is a historical accident.
How was agreement created?
Language change still holds mysteries. If children learn from their parents, where does language
change occur? There are external and internal reasons. Internal language change is when a speaker
of a language fundamentally changes that language. These changes are important for agreement.
Verb second property
Verbs in English always follow the subject. In Dutch, the verb always takes second position. English is
the only Germanic language not to have the property V2. The pattern in a V2 language is as follows:
… V Su pronoun…
When the subject is phonologically reduced, the speakers may feel the need to add another subject,
which creates a pattern like this:
… V Supronoun Su …
This does not agree with Theta-theory. There is one subject without a Theta-role. Only one subject is
interpretable, then is basically becomes agreement.
… V Agr Su …
uPhi: 3SG Phi: 3SG
Negation in French
French has a double negation. The ‘ne’ was introduced and reduced, so people felt the nee to add
“pas”. Now, “pas” is taking over. So “ne” is used for agreement, it is uninterpretable.
Brabantish
Brabantish is a Dutch dialect.
- Hedde da gezien? -de can be analysed as subject.
- Hedde gaai da gezien? ‘gaai’ is now a subject, but what is -de? Agreement!
This also counts in the following sentence:
- Ik geloof da gaai da gezien het (not hedde as there is no agreement in the relative clause!!)
,How is agreement lost?
Middle English Modern English
Over time, English has deflected. The 3rd person SG ending
SG PL SG PL
is now an -s and there is no 2nd person SG ending
1 0 0 1 0 0
anymore. But how did this happen? 2 st 0 2 0 0
It has to do with the pronouns. Speakers felt that “du” 3 th 0 3 s 0
was impolite, so they started using “you” as a 2 nd person
polite form. “You” then completely replaced “du”. The 0-ending goes along with that.
The same story works for Dutch. First it was just like German. “Gij” was used as the polite form and
“du” disappeared. So, the 2nd person SG got a -t ending. Dutch people felt the ambiguity to be weird.
So “gij” (plural) was replaced by “jullie” and “gij” became “jij”,
But why then did the same not happen in Romance languages? In French, “vous” is polite, but French
has not lost “tu”. In Romanian, this is also the case.
The “politeness trigger” is only one factor, the second factor is economy. Why did “Du” and “You”
not co-exist? The answer is that this paradigm is a simpler one. The Modern Romanian
same is true for Dutch, the paradigm becomes simpler as Dutch ended SG PL
up with only 2 affixes instead of three. This factor is known as economy. 1 cânt Cânt-ăm
In Romanian, there would still be 3 different affixes in the SG form. 2 Cânț-i Cânț-ați
This remains so when the 2nd person plural would be used in the 3 Cânt-ă Cânt-ă
singular.
So, there are two factors that play a role for agreement:
1. A social-linguist factor: Politeness
2. Linguist factor: Economy
, Lecture week 2: Constraints on syntactic dependencies
This is a synthesizing lecture. We look at what we have done, and we look further than that.
Language change
1. The [F]-[uF] format
2. C-Command
3. The [F]>[uF] format (which includes c-command
4. A problem
The [F]-[uF] format
There are three syntactic properties:
- Agreement dependency, where the subject triggers the verb
- (accusative) Case dependency, where verbs and prepositions assign accusative case
- (nominative) Case dependency, where a Finhead assigns nominative case to its subject
Two conclusions are that:
- On the surface, these dependencies all look quite distinct.
- There is an adjacency component for the accusative case dependency.
The goal here is to argue that all these dependencies are the same. This is partly already achieved by
describing all the dependencies in the same terms, [F] and [uF]. The interpretable counterpart has no
meaning, in agreement. In the nominative case, the verb [F] assigns case to the subject [uF]. For
accusative case, the verb of preposition [V or P] assigns accusative case [uV or uP].
So, the same principle works for all these cases. You look at the principles behind the constructions
and you can achieve unification.
There is a fourth dependency: Reflexive pronouns
- A personal pronoun can survive in a sentence without an antecedent in that sentence
- A reflexive pronoun cannot.
o “Herself” would not be able to find an antecedent in the context. There is some kind
of restriction.
This seems like a dependency.
Mary (Phi: 3SG fem) loves herself (uPhi: 3SG, fem)
*John (Phi: 3SG masc) loves herself (uPhi: 3SG, fem)
Herself in the second sentence cannot find an interpretable counterpart, and, thus that sentence is
ungrammatical.
C-command
Mary (Phi: 3SG fem) loves herself (uPhi: 3SG, fem)
*Mary (Phi: 3SG fem)’s husband loves herself (uPhi: 3SG, fem)
Why does “herself” need to refer to the whole subject and not to a part of it? It is not enough for a
reflexive to just have an interpretable counterpart. The missing link here is c-command.
C-command can be defined as “A c-commands B is and only if the node that immediately dominates
A also dominates B, and A does not dominate B”. A simpler explanation is that “A c-commands B if
you can reach B from A by going one up and as many as you like down (non-zero) without going
back”.
