Advanced Syntax
description
Transcript of Advanced Syntax
Lecture 7: Tense and Negation
The clause is made up of distinct structural areas with different semantic purposes The VP
One or more verbal head introducing arguments
The IP An inflectional head
introducing finiteness The CP
A complementiser introducing force
Between the thematic VP and the IP there can be other elements which are verbal but not thematic They don’t introduce arguments
These elements are to do with voice and aspect ... That John has been being followed perfect progressive passive
We have argued that these elements are represented by the bound morphemes The verb moves to support the nearest one of these Auxiliaries are inserted to support the rest
-s -en -ing -enshow
be
be
ha
We will distinguish these non-thematic verbal elements by denoting them as “v” (as opposed to V) This is commonly called ‘little v’
While little v is clearly verbal and non-thematic, we cannot say that it is functional either, as that would make it indistinguishable from inflection: Verb = [-F, +V, -N] Inflection = [+F, +V, -N]
We don’t want to add another categorial feature as that would predict an extra 8 more distinct categories
The solution is to allow categories to be undefined on certain features Little v = [+V, -N]
The inflection seems to be able to have a VP or a vP complement, depending on whether there are aspectual morphemes: He may [VP win the race] He had [vP been winning the race]
I selects for [-F, +V, -N] complements This covers thematic VPs which
have exactly these features And vPs, which are not defined for
the [F] feature
Auxiliary verbs are inserted to support verbal morphemes which cannot be supported by the verb
The verb can support at most one overt morpheme He -ed [VP ] He -ed -en [VP the paper] He -ed -ing [VP the paper]
smile se
eread
Auxiliary verbs which follow modals and the infinitive to, do not appear to support any morpheme: They are always in their base form
He may be winning ... for him to have won
If there is no morpheme here, there is no explanation as to why the auxiliary is present
By this reasoning, there must be a morpheme present in these cases
This morpheme is phonologically null He may be- winning ... for him to have- won
It is apparently absent in the presence of a tense inflection: He be–ed (=was) winning * He was be- winning
This might suggest that the morpheme has something to do with tense
Modals and the tense morphemes are in complementary distribution He smiled He will smile * he will smiled
But this is not completely true
Most modals have two forms can could may might shall should will would
Historically these are tense forms In modern English these forms are not usually
used to express tense distinctions He might come he may come (in the past)
But in some cases they are: When I was young, I could play the piano
Modals in subordinate clauses are used to match the tense of the main clause: I think he will come I thought he would come * I thought he will come
This is called tense sequencing
This suggests that tense is not absent with modals
One might have thought that tense and infinitives are mutually exclusive
However, again the situation is not so straightforward
Infinitives are always subordinate Like tense sequencing in finite clauses, the
time reference of the infinitive is often dependent on the main clause
Different types of infinitive have different time interpretations
The temporal interpretation of a control clause is co-temporal with the main clause He tried [PRO to start the car] He is trying [PRO to start the car] He will try [PRO to start the car]
Clearly the ‘trying’ and the ‘starting’ are happening at the same time (regardless of whether it is successful)
Raising clauses are like control clauses in that their temporal interpretation is co-temporal with the main clause The door1 seemed [ t1 to close by itself] The door1 seems [ t1 to be closing by itself] The door1 seems [ t1 to close by itself]
Like tensed clauses the use of the progressive indicates present
time the use of the plain verb indicates the habitual
Exceptional clauses often refer to an event which takes place at some point after that referred to by the exceptional verb I expected [the door to open] I expect [the door to open]
The door opening (or not) happens at some time after the expectation is formed I expected the door to open, but when I tried
it, it was locked
If time reference is mediated through a tense element, it seems that a tense element is present in the infinitive Though not of the same nature as that in finite
clauses In infinitives the tense element is similar to a
pronoun in that its reference is linked to its antecedent i.e. The tense of the main clause
We might refer to this as ‘anaphoric’ tense
This all suggests that tense and inflection are not the same element in a structure
Therefore we would expect them to have different positions
It could be that tense is the phonologically null morpheme that follows modals and the infinitival marker: He may be-pres singing ... for him to be-ana singing
Clearly the tense element follows the inflection
So it must head a phrase that sits in the complement position of the inflection
We know that inflection subcategorises for a [+V, -N] complement
Therefore tense must be of the category ‘little v’
Tensed clauses are still IPs So there must be an inflection when
there is a morphological tense The most obvious suggestion would be
that the inflection is a null morpheme in this case: [IP He [vP -ed [VP smile]]]
This is what was assumed in BESE
But this raises the question of why inflection must be null when tense is overt and tense must be null when inflection is overt
There seems to be a conspiracy to make it look like tense and inflection are in complementary distribution when they are not!
Another possibility is that the ‘tense’ morphemes ARE inflections
Therefore they cannot be ‘tense’ They would get their tense interpretations
from the null tense element that accompanies them [IP He –ed [vP - [VP smile]]]
In this way the tense morpheme is exactly like a modal Though it carries no modal meaning
This allows a very uniform analysis: Tense in English is always null Some inflections select present tense
will, can, shall, may, -s/- Some inflections select past tense
would, could, should, might, must, -ed One inflection select anaphoric tense
to
The fact that an auxiliary must be inserted after a modal and an infinitive in cases where the verb is supporting an aspectual morpheme shows that tense is a bound morpheme:
He will - -ing [VP ]readb
e
The fact that no auxiliary is inserted when there is no aspectual morpheme shows that the verb can move to support tense
He will - [VP ]read
be-
When there is a bound inflection (-ed or –s) and the verb cannot support it, we can assume that the auxiliary is inserted into tense and then moves to support the inflection A verb can support only one OVERT
morpheme, but any number of covert ones
He will -ed -ing [VP ]readb
e-
When there is a bound inflection (-ed or –s) and there is no other over morpheme, the verb can move to support both tense and the inflection
He -ed [VP ]read-
read-
Tense is always null It heads a vP which is the
complement of the inflection Different inflections select for
different tenses Tense is a bound morpheme
which needs supporting When the verb cannot do
this, an auxiliary is inserted Whatever supports tense will
support the bound inflection by moving from v to I
Negation in English is typically marked by the use of not
This sits behind the inflection and in front of the main verb You shall not pass
It can sit anywhere between these two (with meaning differences), but not in front of I or behind V * He not will have been being followed He will not have been being followed He will have not been being followed He will have been not being followed He will have been being not followed * He will have been being followed not
It has been argued that not must be a head of a phrase which sits in the non-thematic verbal part of the structure
This is because it blocks V movement to I He –d [VP smile] => he smile-d
[VP t]
Normally the verb can move to I (via tense) He smile1-2-ed [vP t2 [VP t1 ]]
But when the negative particle is present, this movement is blocked * he smile1-2-ed [ not [vP t2 [VP t1 ]]
This can be explained by Relativised Minimality A head must move to its nearest
head position But this only works if the
negative is a head
In this case, the verb can support the tense
But do has to be inserted to support the inflection He do-ed [ not [ smile1-
[ t1 ]]]
do
If this analysis is correct, the category of the negative particle must be ‘little v’
Because: I selects for a vP complement The phrase headed by the negative can be
the complement of I
Although not does not have many verbal qualities in English, in other languages the negative particle can behave like a verb Finnish
tiedän en tiedäI-know I-not know
tiedät et tiedäyou-know you-not know
The head status of the negative seems to be supported by negative contraction: I haven’t seen him He isn’t here
This might be treated in the same way that bound morphemes are: The auxiliary moves into the negative head
before moving to I He –s –n’t be- here
As heads can only move to head positions, this seems to show that negation is a head
But this predicts the wrong morpheme order as the negation should be closer to the verb than the inflection: * He be-n’t-s here
It also isn’t clear why main verbs cannot do the same thing: * He read-n’t the book
Modal auxiliaries can also bear the contracted negation: I wouldn’t know
But modals cannot move through the negation as they are higher in the structure
The contracted modals sometimes have a different base form Can’t won’t shan’t
Some modals cannot appear with the contracted negation at all *mayn’t
Such observations argue that the contracted negation is better seen as a form of auxiliaries rather than as an independent morpheme
When a bound morpheme cannot be supported by the main verb, an auxiliary is inserted
Which auxiliary is used depends on the following verbal morpheme Main verb do
-ed you know did you know Perfect have
He –ed –en see he had seen Everything else be
He –ed –ing run he was running He –ed –en beat he was beaten
However, the choice of the auxiliary is never dependent on the presence of the negation He did not know me He had not seen me He is not running He was not beaten
In these cases, the negation behaves as though it is invisible It is the verbal element following the negation
that determines which auxiliary to use But this is unexpected if negation is a verbal
head
In general, modifiers do not interfere with syntactic processes: Did he always know the answers? He has sometimes watched the news He is often running
This might suggest that the negative particle is an adverbial modifier rather than a verbal head
So why does it block head movement?
Negative adverbial modifiers have a restricted position with respect to the verb I never win * I win never
This is not so for all adverbial modifiers I sometimes win I win sometimes
We know that the negative particle never follows the verb * he smiled not
When the negative adverb precedes the finite verb, it must precede the inflection which the verb is supporting He never win-s
Negative adverbs can precede inflections (with special emphasis) He never will find out
The negative particle cannot precede the inflection – even with special emphasis * he not will find out
It seems that there are two ordering conditions The negative particle must precede the main verb The negative particle must follow the inflection
These two together will prevent the main verb from moving to I * he not smiled * he smiled not
The only option left is for the verb to stay below the inflection and for the inflection to be supported by do He did not smile
Possibly the best analysis for the negative particle is as an adverbial modifier
There are two conditions on its placement It cannot follow the verb It cannot precede the inflection
Otherwise it can go anywhere between the two