Introduction to General Linguistics WS12/13 page 1 Syntax 6 The finite verb and the clause: Course teacher: Sam Featherston Important things you will learn in this section: The head of the clause The positions of modals and auxiliaries in English and German The position of finite verbs in English and German. The third schema for building phrase structure: specifiers The subject position as specifier of 1 is the head of a clause Above a VP there is another projection, the. The I in stands for Inflection. The is basically verbal in the same sense that a is basically nominal. The is a functional verbal projection over the lexical VP, in the same way that the is a functional nominal projection over the lexical NP. (1) I VP The defining core of a clause is its verbal features of finiteness, person, number, and tense. These are often marked on the finite verb. All clauses can be finite (he could play football) or non-finite (... him to play football). Since finiteness is a feature of the whole clause, it marked on the head of the clause. This explains why a finite clause must have one finite verb. (2) I[+fin] VP could play football Similarly, a non-finite clause must thus have exactly one non-finite verb in the infinitive. The word to marks a verb as infinitive and is thought to appear in I[-finite]. A whole clause is thus a category. (3) But where is the subject? I[-fin] VP to play football The subject is also within the projection, but it is neither a complement nor an adjunct, but a specifier. We will discuss this below. The subject is in the specifier position of, daughter of and sister of I' ( I-bar ), which is a new level of projection between I and. A full finite clause therefore has this structure: (4)
Introduction to General Linguistics WS12/13 page 2 the mouse I[+fin] VP could V see the cheese 2. The specifier position We have three basic phrase structure components so far: head, complement, and adjunct. The head is the X which projects into an XP. A complement is sister to an X. The order of head and complement depends upon the head type and the language, but it is usually fixed. (5) XP XP X ZP ZP X head complement complement head An adjunct is sister to an XP and daughter to an XP. Adjuncts can often appear either side. (6) XP XP YP XP XP YP adjunct host phrase host phrase adjunct We now add the last one: the specifier. The specifier, a phrase YP, is daughter of XP and sister to an intermediate projection called X-bar, and shown as X or X'. Specifiers appear on the left. (7) XP XP YP X' YP X' specifier specifier X ZP ZP X head complement complement head Specifiers are usually only on the left. (Unlike complements and adjuncts.) The complement is in an internal position: sister to X, the head The adjunct is in a outside position, sister to XP and daughter of an extra XP. The specifier is in a middle position: sister to X', a level of projection between X and XP. They are daughters to XP, the maximal projection. The specifier is a phrasal position. When there is no specifier, syntacticians usually leave out the X'. A short name for specifiers is spec. Linguists talk about spec- or SPEC, as the specifier of. The subject, in spec-, and the finite verb, in the I position, are in a spec-head relation. This is typically one of agreement. In many languages, finite verbs and subjects agree in their feature specifications of person, number, and gender.
Introduction to General Linguistics WS12/13 page 3 3 The syntax of the I position 3.1 Evidence for I position from modals Why do we think there is a category I and a position above the VP? In English, there is evidence that the modal verbs are of this category I[+fin], and always occur in this position. (8) the mouse I VP could V will might see the cheese The evidence can be seen when we look at the way English modals differ from normal verbs. (9) English modals: only one: * He will (to) must... * He should (to) can... (10) English modals: never under auxiliary: * He has must / musted... * She is can / canning (11) English modals: never in infinitives: * to must do something * to can do something Speakers of English want to say these things, but the language doesn't allow them. They have to use have to, be able to, be going to... Why doesn't the language allow these forms of the modals? Analysis: normal verbs are V, heading a VP. They can take auxiliaries, and thus appear in compound tenses like the perfect (have done), the progressive (be doing), and the passive (be done). They can also appear as infinitives (to do). They can be optionally be finite, and carry the [+finiteness] feature, or non-finite. English modals are not of the category V, but of the category I and must carry finiteness, the feature [+finite]. Since a sentence can only have one head, it can only have one I, which means that modals cannot be complements of another higher verb. The [+finite] feature excludes the occurrence as an infinitive.
Introduction to General Linguistics WS12/13 page 4 3.2 Evidence for I position from negation Negation not follows modals. The analysis we adopt here places negation as an adjunct to I. (12) the mouse I VP I neg V will not see the cheese This makes it natural that some modals can fuse with negation to form a single word. (13) should + not = shouldn't will + not = won't On the other hand, a finite main verb cannot be in I. Thus it cannot precede not : (14) * She sees not him the mouse I VP I neg V *see-s not Ø the cheese Nor can it fuse with the negative not, which also suggests that normal verbs cannot appear in I. (15)* seesn't * seen't * seen'ts Negative clauses without modals require do-insertion: *She sees not him. *She not sees him. (16) She does not see him the mouse I VP I neg V do-es not see the cheese Why is the form of do required? The auxiliary do does not add any meaning here. It seems that do provides the head of the sentence I with something to put its inflection on. Notice that this dummy do can also fuse with the negative not, like modals, which we think are in I. (17) don't doesn't
Introduction to General Linguistics WS12/13 page 5 3.3 Evidence for the I position from adverbs Adverbs provide evidence for the I position too. Adjuncts appear at the edges of VPs... (18) on the left... VP or on the right... VP AP VP VP AP A V V A quickly eat a pizza eat a pizza quickly So an adverb such as quickly can appear at either end of a VP, but not inside the VP or inside the. Notice that finite auxiliaries usually pattern the same way as modals. (19) a. I must [ VP quickly eat a pizza]. I have [ VP often eaten Indian food]. b.*i quickly must [ VP eat a pizza]. *I often [ VP have eaten Indian food]. c.*i must [ VP eat quickly a pizza]. *I have [ VP eaten often Indian food]. d. I must [ VP eat a pizza quickly]. I have [ VP eaten Indian food often]. To see this more clearly, we show the sentences in more detail: (20) a. [ I (*quickly) must [ VP (quickly) eat (*quickly) a pizza (quickly) VP ] ] b. [ I (*often) have [ VP (often) eaten (*often) Indian food (often) VP ] ] Note that the position of the adverb is not just "after the first verb", because the adverb stays in the same place even if there is no modal. Compare (19)a-d and (22)a,c,d to see this: (21) a. I quickly eat a pizza. I often cook Indian food. c.*i eat quickly a pizza. *I cook often Indian food. d. I eat a pizza quickly. I cook Indian food often. 4 Auxiliaries can move to I Auxiliaries seem to occupy the I position too (see above I have often eaten Indian food) but they have infinitive forms (to have seen him) and appear as the complement of other verbs (I should have seen him). It seems that auxiliaries are of the category V, and start off in their own V position, but can move to the position I, if they are finite. (22) the mouse I VP ha-s V VP seen V the cheese Support for this is the fact that adverbs follow finite auxiliaries as they follow modals. (23) a. I have often eaten Indian food. perfect have b. I am always eating pizza. progressive be c. I am rarely found in my office. passive be
Introduction to General Linguistics WS12/13 page 6 The negative not follows auxiliaries too and can fuse with them. (24) a. I haven't eaten Indian food. perfect have b. She isn't eating pizza. progressive be c. He isn't found in his office. passive be But we can see that they seem to stay in their V base position, if the I is not empty. (25) She must often have eaten Indian food. 5 Main verbs and I: Trouble Clauses without auxiliaries or modals are still centred around the, which is still specified as [+finite]. But the finite affix -s in this structure. does not appear in the I position, but is attached to the verb in V. The reasons for this are complex and poorly understood, but it seems that there is an 'abstract relation' between the I position and the inflection of the verb. Sometimes is it said that the features move down from the I to the V position to meet the finite verb. Sometimes it is thought that the verb moves up 'covertly'. We need not solve this problem here. (26) the mouse I[+fin] VP V see-s the cheese We still cannot put an adverb after a main verb, even if it carries the finite inflection, for instance. Note that this inflection may be a zero morpheme -. (27) a. She quickly eats/ate a pizza. I often eat- /ate Indian food. b.*she eats/ate quickly a pizza. *I eat- /ate often Indian food. c. She eats/ate a pizza quickly. I eat- /ate Indian food often. Interestingly, a not seems to break the abstract relation between I and the inflectional ending on a finite main verb. It would appear that the not requires something visible to hang on to. Or perhaps the 'abstract link' between the V and the I position requires the to be directly next to each other. (28) * She not sees him the mouse I VP I[+fin] neg V not see-s the cheese
Introduction to General Linguistics WS12/13 page 7 6. in German But what about German? The analysis was developed from English, but the data in German is rather different. German auxiliary and modal verbs do not behave differently from other verbs. They behave the same as normal German verbs, that is: (29) German modals can stack up: English modals: only one:... dass er Erdbeeren kaufen müssen soll * He should must...... dass er das machen können wird * He will can... German modals can be under an auxiliary: English modals: never under auxiliary:... dass er das kaufen müssen hat * He has must... German modals can occur in infinitives: etwas machen (zu) müssen English modals: never in infinitives: * to must do something German modals: have V-morphology: English modals: no inflectional morphology: Ich rufe, du ruf-st,... I sing, she sings,... Ich kann, du kann-st,... but: I can, she can,... Analysis: German modals are verbs, heading a VP. Their inflection is like the inflection on other verbs, including non-finite inflection. Like other verbs, they can be embedded. English modals are not of the category V, but of the category I[+fin]. German is simpler than English. There are two major different analyses of the in German. The first analysis: German has an. The is head-final, like the German VP. (30) (weil) du die Pizza gegessen hast because you the pizza eaten have "...because you have eaten the pizza" VP I[+fin] VP V ha-st V die Pizza gegessen This structure shows that hast, the perfect auxiliary, is of the category V, but can move to the finite I position. On this analysis, German main verbs and auxiliaries are of the V category just as in English, and German has an just as in English. The problem with this analysis is that there is very little evidence *for* an in German. Adverbs and the negative nicht never appear between the two, and German modal verbs behave almost exactly the same as main verbs. Is it not possible that the apparent I position in English is simply a strange piece of behaviour of English, and not general to all languages? For this reasons, many German linguists prefer the second analysis that German has no, but just a VP which can have the feature [+ finite]. (31) VP
Introduction to General Linguistics WS12/13 page 8 VP V[+fin] V ha-st die Pizza gegessen On the other hand, there is little evidence *against* the existence of an in German. Because the two head positions V and I are always just immediately adjacent, we would not expect the evidence that we find in English to be visible in German. Perhaps German has an, but simply has no modals which fill it. There is a trend in this direction in English too. Verbs such as ought, dare and need used to be modals, but are rapidly becoming normal verbs. (32) Old Need you smoke cigars in bed? Dare you criticize her dress? New Do you need to smoke cigars in bed? Do you dare criticize her dress? NB We shall assume that there is an in German. The subject is in spec- in German just as in English. Our structure of an embedded German clause therefore looks like these examples: (33) (weil) ihr die Pizza essen wollt because you the pizza eat wants 'because you want to eat the pizza' I' ihr VP I[+fin] VP V woll-t V die Pizza essen (34) (weil) die Maus den Käse gesehen hat because the mouse the cheese seen has 'because the mouse has seen the cheese' I' die Maus VP I[+fin] VP V ha-t V den Käse gesehen Exercises on in German and English 1. She must work 2. I can't sing 3. We shouldn't wait 4. Sam will like pizza 5. He has come 6. He is reading
Introduction to General Linguistics WS12/13 page 9 7. I may find it 8. He is not reading 9. She smokes 10. He doesn't smoke 11. She hasn't visited 12. He runs slowly 13. I can't hate him really. 14. (I want) him to believe me. 1. (weil) ich gekommen bin. 2. (weil) ich rauche because I come am because I smoke 'because I have come' 'because I smoke' 3. (weil) ich es gekauft habe. 4. (obwohl) ich ihn liebe because I it bought have although I him love 'because I have bought it' 'although I love him' 5. (dass) ich die Schule verlassen haben sollte 6. (ob) mein Motorrad schnell fährt that I the school left have should whether my motorbike fast goes 'that I should have left the school' "whether my motorbike goes fast" Homework 1. The kind manager of the hotel 2. I will eat pizza. 3 I have not eaten it 4. I drink milk 5. I don't drink milk 6. I often drink it. 7. I don't buy it often. 8. I should always drink it 9. (I want) him to apologize politely 10. My house smells. 11. (weil) sie langsam gearbeitet hat 12. (weil) ich oft gesehen worden bin she slowly worked has I often seen been am 13. (dass) der Knabe den Drachen töten kann 14. (dass) der Käse mir geschmeckt hat the boy the drageon kill can the cheese me tasted.nice has 15. (weil) du es getrunken hast 16. (ob) die Maus den Faden abbeißt you it drunk have the mouse the thread bites.off
Introduction to General Linguistics WS12/13 page 10 Rules so far... 1. All lexical items belong to one of these categories A, N, V, D, I, (also C 'complementizer') 2. All categories project into a phrasal category. So for instance N is the head of NP. 3. All categories can have complements. Complements are sisters of the head (eg N) and daughters of the phrasal projection (eg NP). Complements are phrasal projections (eg ). (i) XP X YP 4. In English, complements are always on the right. In German, the categories D, N, and C take their complements on the right: D: der Hund ('the dog') N: Mutter von dem Kind ('mother of the child') C: dass er kommt ('that he comes') The categories V and I (if it exists in German), take their complements on the left: V: die Kinder lieben ('love the children') I: geliebt hat ('has loved') The category A can take its complement either side: A: auf die Kinder stolz, stolz auf die Kinder ('proud of the children'). 5. We must distinguish arguments and adjuncts. Arguments are participants in an action or situation which are already part of the meaning. (eg wife means 'female partner of another person', so this other person is an argument; give has three arguments, the giver, the thing given, and the person receiving it.). Adjuncts are just optional addition details. 6. Most arguments (but not subject arguments) are complements, so sisters to X. Adjuncts are attached to phrasal projections 'on the outside', at least in English. They are thus sisters to XP and daughters to an additional XP which we add on top. XP XP X YP XP ZP complement adjunct 7. Certain sorts of verbs can take a VP as their complement. In English, the perfect auxiliary verb have takes a VP with the status past participle, the progressive auxiliary verb be takes a VP in the present participle form, passive auxiliary verb be takes a VP as past participle. Modals take bare infinitives. These verbs can therefore stack. English: VP 1 German: VP 1 V 1 VP 2 VP 2 V 1
Introduction to General Linguistics WS12/13 page 11 8. Above the VP there is a verbal functional category. In English: - modals can only appear in I, - finite auxiliary verbs move from their own V position into I, - other verbs (non-finite auxiliaries and main verbs) never move to I. We are forced to assume an 'abstract relation' between finite main verbs and the finiteness feature in I (eg She [ I [+fin] ] see-s her mother) to account for this. 9. The negative not is assumed to be an adjunct to I. ('head adjunction') 10. German is much simpler: all verbs behave the same. All verbs are start in V; the one finite verb moves from its V to I. We therefore analyze English modals and dummy do as being in I. A not may be adjoined to the I position. D I VP she I neg V will not see D might does him Auxiliaries move into I to carry the finiteness feature, if they are the top verb, and thus need to be finite. D I[+fin] VP she ha-s V VP V seen D him Main verbs do not move to I, but stand in an 'abstract relation' to I and the finiteness feature. D I[+fin] VP she V see-s D him
Introduction to General Linguistics WS12/13 page 12 All German verbs are V, none start in I. The top verb moves to I to carry the finiteness feature.... weil du die Pizza gegessen hast I' du VP I[+fin] VP V ha-st V die Pizza gegessen... weil wir die Pizza gekauft haben sollten VP V V haben die Pizza gekauft I' wir VP I[+fin] VP V sollt-en 10. All German VPs are head-final, so their complements appear on the left. German s are head-final too, and so their complement VP appears on the left too. But subject are specifiers. These also appear on the left, just like VP specifiers in English. The main difference between English and German so far is that English verbal projections and VP are head-initial, while German verbal projections are head-final. This simple difference alone explains much of the contrasting clause structures in English and German. It is therefore a verb good example of how generative grammar can explain things, not just describe them.