On the extension of the X-bar principles to the functional categories such as inflections, complementisers and determiners
Mark Newson
AN-341
5th April 1994
On the extension of the X-bar principles
to the functional categories such as
inflections, complementisers and determiners
by Peter Gelleri
ELTE BTK DELL 1994
"There are three things in life you must never run after: a woman, a bus, and a theory of transformational grammar - there will be another one along in a moment"
(A well-known linguist, Aitchison 1978)
On the extension of the X-bar principles to the functional categories such as inflections, complementisers, and determiners
Getting generalised
Having waded through various essays and lecture notes on X-bar Theory, one cannot avoid noticing the strikingly frequent occurrences of the words "universal", "general" and "generalised". There is no doubt about the fact that Noam Chomsky, and other linguists under the influence of the Chomskian theories and that of what is usually referred to as modern American linguistics, tried to work out a system which would fit all human languages (Aitchison 1978), consequently this grammatical framework must be a really general one.
(Newson, 1994:lec.5.,9.; slightly modified):
(1)
XP = X'' (maximal projection)
?P X' X = either N, A, Adv, or P
? = N, A, Adv, P
X ?P
The canonic X-bar tree diagram (1) appears to be a highly generalised structure which can easily be expanded into larger units (2). All we have to do is duplicate the basic structure and attach the elements one after the other. Hence the assumption that language acquisition is based upon one single structure (complement rule and specifier rule), and using the X-bar Theory we can analyse and (re)generate constructions in language from word-level up to the larger units of phrases and even sentences.
(2)
?P
(?P) ?'
? ?P
(?P) ?'
? ?P
In spite of its simplicity and general nature, the X-bar Theory is said to be efficient regardless of what language we choose for analysis. Our task, however, is not to examine whether the Theory works with all human languages but rather to see if/how it can handle functional categories such as inflections, complementisers, and determiners.
INFL
As far as inflections are concerned, one major advancement of Transformational Grammar is the separation of auxiliary verb (AUX) from the verb phrase (VP) and the formulation of a new node labelled as INFL (inflection). This can be done through Deep Structure Analysis (Aitchison, 1978). (3b) shows the DS of (3a):
(3a) Noam is out of his mind.
(3b) Noam PRES + be out of Noam's mind.
We can include the new element, I, in the X-bar analysis, which proves both the validity of the Theory and the realisation of INFL as a separate node (Haegeman, 1991:100,105). Thus we can fully apply the X-bar principles to I and draw our diagram for (3a):
(3c)
IP
NP I'
Noam
I VP
PRES
3/sing be out of his mind
Although this structure analysis appears to be well-suited in all cases, one trying to diagnose (4a) may find oneself baffled.
(4a) Noam should have been a greengrocer.
At first sight there seems to be no problem with (4a) at all. VP comprises of the primary auxiliary be plus the NP, a greengrocer. Noam as a NP takes the specifier position for I-bar.
(4b)
IP
NP I'
Noam
I VP
should
have + pastP be a greengrocer
As illustrated in (4b), there are two auxiliaries (a modal and a primary one) under I. The problem arises when we form a question. According to the I-movement rule (Radford, 1988; "Head-to-Head movement", Haegeman, 1991) I should be moved and raised to the position of C that specifies for IP (I-double-bar), i.e. I is placed before the subject NP. Still if we apply this rule, the result will be an incorrect English sentence:
(5) * Should have Noam been a greengrocer?
Thus we have to propose that it is only the first auxiliary (operator) that gets moved out of I, consequently in case of more than one auxiliary verbs, I remains to be occupied. The fact that the operator gets moved while the primary auxiliary stays under I and cannot be displaced indicates that bringing them together under the same bottommost node is not a perfect solution of the IP analysis (Moreover, a sentence may have a long string of verbs denoting tense, person, mood, and aspect; eg.: Noam may have been thinking about being a greengrocer)
To resolve this problem and preserve the validity of the I-movement rule, we could adopt the adjunct rule (Newson 1994:lec.5,12) and
the corresponding rewrite rule ( X' -> X' YP ). Installing a new node, say AuxP, under which the stationary primary auxiliary is placed, we get the following diagram:
(6)
IP
NP I'
I'
I AuxP VP
Noam should have been a greengrocer.
0
COMPL
It seems plausible to posit that complementisers (C) take an IP complement (Haegeman, 1991:106). Thus we can extend the X-bar theory to clauses introduced by a C element such as that, whether, for, or if). Let us now consider the following examples:
(7a) Noam thinks that trees are nice.
(7b) * Noam thinks that whether trees are nice.
(7c) Noam thinks that whether trees are nice is unimportant.
Since C-node cannot take more than one complementisers, (7b) is ungrammatical (Haegeman, 1991). On the other hand, (7c) is a correct English sentence although it contains two complementisers next to each other. It is time we turned to the X-bar analysis again.
((8a) and (8c) are diagrams for (7a) and (7c) respecitvely.)
(8a)
CP
C'
C IP
that trees are nice
(8c)
CP
C'
C IP
that
CP I'
whether trees are nice is unimportant
We have come to the conclusion that the X-bar format can also be applied to more complex structures containing C element(s).
DET
To attest whether or not the functional category of determiners fits into, and can be tackled by, the X-bar framework, let us first consider (9) below.
(9a) Noam drives a Trabant.
It is superficial that the indefinite article in (8) belongs to the NP it precedes. (It cannot be separated from its complement). An X-bar analysis verifies this assumption and reveals the relationship between the determiner and its NP.
(9b)
DP
D'
D NP
a Trabi
The X-bar diagram of (9b) above clearly exemplifies that the determiner takes the subsequent NP as its complement and gets projected into D-bar.
But things are a bit more complicated than they seem. Our next example is as follows:
(10a) Noam fails his every student.
(10b)
DP
D'
D DP
D'
D NP
his every student
If we try and generate other sentences that conform to the pattern of (10b), we are likely to get monstrosities such as (11a) and (11b).
(11a) * Noam hates this that book.
(11b) * Noam sold some his houses.
Although every may co-occur with possessives (Quirk), and therefore (10b) is possible, in this case as an exception the X-bar analysis has apparently failed to provide us with a general structure.
We may, however, find the Theory applicable and versatile enough when analysing of-constructions of determiners (12a). Our diagram (12b), though, will look somewhat different from (10b).
(12a) Noam sold some of his houses.
(12b)
DP
D'
D NP
some of his
(see: Newson, 1994:lec.5.,2/b)
CONCL
The conclusion of our brief survey could be that even though we have an extensively generalised pattern at our disposal, we might still need certain particularisations and subcategorisations in order to fine-tune the system. As we said earlier, things are always a little bit more complicated than they seem.
Bibliographical Notes
Haegeman, Liliane. Introduction to Government and Binding Theory Basil Blackwell, 1991.
Radford, Andrew. Transformational Grammar (A first Course); chpt. 9.10-9.13. Cambridge Univ.P., Cambridge, 1988.
Newson, Mark. Lecture notes (AN-341)
ELTE BTK DELL Budapest, 1994.
Aitchison, Jean. Linguistics Teach Yourself Books,
Hodder and Stoughton, 1978.
Andras, L.T.-Varga L. ed. Introductory Readings in Modern Linguistics
Tankonyvkiado, Budapest, 1983.
Quirk - Greenbaum. A University Grammar of English
Longman, (1973) 1987.
AN-341
5th April 1994
On the extension of the X-bar principles
to the functional categories such as
inflections, complementisers and determiners
by Peter Gelleri
ELTE BTK DELL 1994
"There are three things in life you must never run after: a woman, a bus, and a theory of transformational grammar - there will be another one along in a moment"
(A well-known linguist, Aitchison 1978)
On the extension of the X-bar principles to the functional categories such as inflections, complementisers, and determiners
Getting generalised
Having waded through various essays and lecture notes on X-bar Theory, one cannot avoid noticing the strikingly frequent occurrences of the words "universal", "general" and "generalised". There is no doubt about the fact that Noam Chomsky, and other linguists under the influence of the Chomskian theories and that of what is usually referred to as modern American linguistics, tried to work out a system which would fit all human languages (Aitchison 1978), consequently this grammatical framework must be a really general one.
(Newson, 1994:lec.5.,9.; slightly modified):
(1)
XP = X'' (maximal projection)
?P X' X = either N, A, Adv, or P
? = N, A, Adv, P
X ?P
The canonic X-bar tree diagram (1) appears to be a highly generalised structure which can easily be expanded into larger units (2). All we have to do is duplicate the basic structure and attach the elements one after the other. Hence the assumption that language acquisition is based upon one single structure (complement rule and specifier rule), and using the X-bar Theory we can analyse and (re)generate constructions in language from word-level up to the larger units of phrases and even sentences.
(2)
?P
(?P) ?'
? ?P
(?P) ?'
? ?P
In spite of its simplicity and general nature, the X-bar Theory is said to be efficient regardless of what language we choose for analysis. Our task, however, is not to examine whether the Theory works with all human languages but rather to see if/how it can handle functional categories such as inflections, complementisers, and determiners.
INFL
As far as inflections are concerned, one major advancement of Transformational Grammar is the separation of auxiliary verb (AUX) from the verb phrase (VP) and the formulation of a new node labelled as INFL (inflection). This can be done through Deep Structure Analysis (Aitchison, 1978). (3b) shows the DS of (3a):
(3a) Noam is out of his mind.
(3b) Noam PRES + be out of Noam's mind.
We can include the new element, I, in the X-bar analysis, which proves both the validity of the Theory and the realisation of INFL as a separate node (Haegeman, 1991:100,105). Thus we can fully apply the X-bar principles to I and draw our diagram for (3a):
(3c)
IP
NP I'
Noam
I VP
PRES
3/sing be out of his mind
Although this structure analysis appears to be well-suited in all cases, one trying to diagnose (4a) may find oneself baffled.
(4a) Noam should have been a greengrocer.
At first sight there seems to be no problem with (4a) at all. VP comprises of the primary auxiliary be plus the NP, a greengrocer. Noam as a NP takes the specifier position for I-bar.
(4b)
IP
NP I'
Noam
I VP
should
have + pastP be a greengrocer
As illustrated in (4b), there are two auxiliaries (a modal and a primary one) under I. The problem arises when we form a question. According to the I-movement rule (Radford, 1988; "Head-to-Head movement", Haegeman, 1991) I should be moved and raised to the position of C that specifies for IP (I-double-bar), i.e. I is placed before the subject NP. Still if we apply this rule, the result will be an incorrect English sentence:
(5) * Should have Noam been a greengrocer?
Thus we have to propose that it is only the first auxiliary (operator) that gets moved out of I, consequently in case of more than one auxiliary verbs, I remains to be occupied. The fact that the operator gets moved while the primary auxiliary stays under I and cannot be displaced indicates that bringing them together under the same bottommost node is not a perfect solution of the IP analysis (Moreover, a sentence may have a long string of verbs denoting tense, person, mood, and aspect; eg.: Noam may have been thinking about being a greengrocer)
To resolve this problem and preserve the validity of the I-movement rule, we could adopt the adjunct rule (Newson 1994:lec.5,12) and
the corresponding rewrite rule ( X' -> X' YP ). Installing a new node, say AuxP, under which the stationary primary auxiliary is placed, we get the following diagram:
(6)
IP
NP I'
I'
I AuxP VP
Noam should have been a greengrocer.
0
COMPL
It seems plausible to posit that complementisers (C) take an IP complement (Haegeman, 1991:106). Thus we can extend the X-bar theory to clauses introduced by a C element such as that, whether, for, or if). Let us now consider the following examples:
(7a) Noam thinks that trees are nice.
(7b) * Noam thinks that whether trees are nice.
(7c) Noam thinks that whether trees are nice is unimportant.
Since C-node cannot take more than one complementisers, (7b) is ungrammatical (Haegeman, 1991). On the other hand, (7c) is a correct English sentence although it contains two complementisers next to each other. It is time we turned to the X-bar analysis again.
((8a) and (8c) are diagrams for (7a) and (7c) respecitvely.)
(8a)
CP
C'
C IP
that trees are nice
(8c)
CP
C'
C IP
that
CP I'
whether trees are nice is unimportant
We have come to the conclusion that the X-bar format can also be applied to more complex structures containing C element(s).
DET
To attest whether or not the functional category of determiners fits into, and can be tackled by, the X-bar framework, let us first consider (9) below.
(9a) Noam drives a Trabant.
It is superficial that the indefinite article in (8) belongs to the NP it precedes. (It cannot be separated from its complement). An X-bar analysis verifies this assumption and reveals the relationship between the determiner and its NP.
(9b)
DP
D'
D NP
a Trabi
The X-bar diagram of (9b) above clearly exemplifies that the determiner takes the subsequent NP as its complement and gets projected into D-bar.
But things are a bit more complicated than they seem. Our next example is as follows:
(10a) Noam fails his every student.
(10b)
DP
D'
D DP
D'
D NP
his every student
If we try and generate other sentences that conform to the pattern of (10b), we are likely to get monstrosities such as (11a) and (11b).
(11a) * Noam hates this that book.
(11b) * Noam sold some his houses.
Although every may co-occur with possessives (Quirk), and therefore (10b) is possible, in this case as an exception the X-bar analysis has apparently failed to provide us with a general structure.
We may, however, find the Theory applicable and versatile enough when analysing of-constructions of determiners (12a). Our diagram (12b), though, will look somewhat different from (10b).
(12a) Noam sold some of his houses.
(12b)
DP
D'
D NP
some of his
(see: Newson, 1994:lec.5.,2/b)
CONCL
The conclusion of our brief survey could be that even though we have an extensively generalised pattern at our disposal, we might still need certain particularisations and subcategorisations in order to fine-tune the system. As we said earlier, things are always a little bit more complicated than they seem.
Bibliographical Notes
Haegeman, Liliane. Introduction to Government and Binding Theory Basil Blackwell, 1991.
Radford, Andrew. Transformational Grammar (A first Course); chpt. 9.10-9.13. Cambridge Univ.P., Cambridge, 1988.
Newson, Mark. Lecture notes (AN-341)
ELTE BTK DELL Budapest, 1994.
Aitchison, Jean. Linguistics Teach Yourself Books,
Hodder and Stoughton, 1978.
Andras, L.T.-Varga L. ed. Introductory Readings in Modern Linguistics
Tankonyvkiado, Budapest, 1983.
Quirk - Greenbaum. A University Grammar of English
Longman, (1973) 1987.