"this big" and "too big"

[Edith A Moravcsik]

On July 7, I posted some questions regarding the ordering of a subtype of
adjectival phrases in English. This is a summary of the responses, most of
which came to me directly rather than to FUNKNET. The respondents were
Michael Darnell, Deborah DuBartell, Richard Hudson, Bingfu Lu, Carl Mills,
Peter, whose last name was not given in his message, and Arnold Zwicky.
Thank you for your contributions!

The original query pertained to the exceptional ordering  of English
adjectival phrases such as "this big" and "too bright". The majority of
other adjectival phrases take an internal position within the noun phrase:
they follow the article and precede the noun, as in

     a big apple
     a very big apple.

This internal position is not available for "this big", "too big" and some
others:

     *a this big apple
     *a too big apple;

instead, they are ordered on the peripheries of the noun phrase, either
preceding the article or, some of them, following the noun:

      this big an apple
      too big an apple
      an apple this big
     *an apple too big.

The question is why these adjectival phrases differ from other adjectival
phrases in this manner.

Below, I will refer to "this big", "too big", etc. as "exceptional
adjectival phrases" and to their positions as "pre-article position" and
"post-noun position". I will summarize the responses in two groups: those
pertaining to data and those pertaining to analysis. The outline of this
report will thus be the following:

   l. Data
      (A) Selection
      (B) Order
      (C) Variation
          (a) The pre-article position
          (b) The post-noun position
   2. Analyses
      (A) The pre-article position
          (a) Lu's account
          (b) Zwicky's account
      (B) The post-noun position
          (a) Other post-nominal adjectival phrases
              in English
          (b) Ellipsis
   3. Conclusions
   References and additional bibliography

l. DATA
   (A) SELECTION
   First, let us consider selectional constraints INTERNAL to exceptional
adjectival phrases. Such phrases consist of an adjective and a modifier of
that adjective. While there do not appear to be any constraints on what
the adjective can be as long as it is semantically compatible wit the
modifier, the possible modifiers form a limited set. Hudson and Zwicky
list several of them. The emerging list is this:
(Zwicky 1995: 114)
      "so...that S",
      "SO(emphatic), that",
      "as...as S/NP",
      "too (...(for NP) to VP)"
      "more (...than NP) (in negative contexts)
      "how", "however"

   Second, let us look at the CONTEXT which which exceptional adjectival
phrases can be placed. The rest of the noun phrase within which they occur
is strictly limited: as pointed out by Hudson and by Zwicky (1995: 115,
118), the article must be "a(n)" and the noun must be singular and
countable. Thus, the definite article is ungrammatical (*"this big the
apple") and so are plural nouns. For example, as noted by Hudson, "How big
a car does your neighbor have?" is OK but "How big cars do the people in
your road have?" is not and this fact presents a communication gap: there
is no simple way to convey the intended meaning.

   Notice, however, that the singular-noun requirement holds for only one
of the two exceptional orders of the adjectival phrases - in particular,
for the pre-article position but not for the post-noun position. Thus, "I
don't like this shiny cars." (where "this" is intended as a modifier of
"shiny" rather than a determiner of "cars") is ungrammatical but "I don't
like cars this shiny." is OK. This suggests that the two exceptional
peripheral positions - pre-article and post-noun - form different
constructions. The loosening of selectional constraints in the case of
postposed adjectival phrases may be a manifestation of the generally less
tight connection between noun and postposed attributes manifested in many
languages.

   In other ways, however, exceptional adjectival phrases are
selectionally like "normal" ones. Zwicky (1995: 116) points out that in
that they can occur predicatively ("The table is very wide." "The table is
this wide.") and in "though-fronting" constructions ("Rather big though
the box was...", "Too big though the box was...") just as other adjectival
phrases can.

   (B) ORDER
   The PHRASE-INTERNAL order in exceptional adjectival phrases is regular:
modifier followed by adjective. It is their order RELATIVE TO THE REST OF
THE NOUN PHRASE that makes them linearly exceptional: as seen above, they
are peripherally, rather than internally, placed within the noun phrase.
The two peripheral positions in question - pre-article and post-noun - are
not on a par. First, as we just saw, the rest of the noun phrase is
selectionally more constrained if the adjectival phrase precedes than when
it follows. Second, not all exceptional adjectival phrases that can occur
in pre-article position can also occur in post-noun position. Thus, "this
Adj" can occur in either position but "too Adj", "how Adj" etc. see to be
restricted to the pre-article position; cf.

   this ornate a building
   a building this ornate

   too ornate a building
  *a building too ornate

   how ornate a building
  *a building how ornate

   Zwicky's judgment differs, however. According to him (1995: 122), "too
Adj" can also be postposed: he gives the example "any candidate too
tired". While a sentence such as "Any candidate too tired can leave the
room." does indeed sound good, "I didn't see any candidate too tired."
sounds much worse, at least to my non-native ears, than "I didn't see any
candidate this tired." Multiple native-speaker judgments would be needed
to decide whether "this Adj" and "too Adj" are equally postposable.

   Thus, the two kinds of adjectival phrases - exceptional ones and (most)
normal ones - form subclasses of a single class in terms of both their
selectional and their linear properties. The two phrase types are
selectionally and linearly alike in their internal sctructure and in that
they can both occur in predicative position and in "though"-fronting"
constructions occupying the same position. However, exceptional adjectival
phrases are linearly different from normal ones in that within the noun
phrase, the regular internal position is not available to them; instead,
they are ordered peripherally. They in turn fall into two selectionally
and linearly significant subclasses: all of them can occur preceding the
article, in which case they must occur with indefinite singular nouns, but
some of them may also follow the noun, in which case these selectional
restrictions are relaxed.

   (C) VARIATION
   (a) THE PRE-ARTICLE POSITION
   The pre-article position of exceptional adjectival phrases is subject
to varying grammaticality judgments.
Carl Mills made two relevant points relative to his home dialect
identified as white, working class, rural, and mostly south midland
American. First, he says exceptional adjectival phrases can show
non-exceptional order in that they may occur in the regular, noun-phrase
internal position. Thus, sentences as "Peter ate a this/too big apple.",
he says, are not likely but nonetheless possible in this dialect.

   His second comment is that, while the pre-article position is possible,
it is equally likely that "of" will intervene between modifier and
adjective. Thus, in addition to "Peter ate this/too big an apple.", there
is the equally likely "Peter ate this/too big of an apple." The
"of"-construction was also noted by Darnell and Hudson and it is discussed
in Zwicky 1995: 113, 116-117 in the context of other "grammatical" (rather
than semantic) uses of "of".
This construction may in fact be a legitimate alternative to the "of-less"
version in all dialects of English.

   In addition to differences in how dialects of English evaluate the
ordering options of exceptional adjectival phrases, it should be noted
that entire languages, too, also differ in this regard. Thus, German "ein
so grosses Klavier" "a so big piano" and "ein zu grosses Klavier" "a too
big piano" are fine (and the English-like orders "so gross(es) ein
Klavier" and "zu gross(es) ein Klavier" are ungrammatical I suspect; but
this needs to be checked). Hungarian, too, orders the semantic equivalents
of the English exceptional adjectival phrases as all others; i.e., after
the article: "egy ilyen nagy zongora" "a so big piano" 'so big a piano'
and "egy tu'l nagy zongora" 'too big a piano' are fine but the
English-type pre-article order of the adjectival phrase, as in "ilyen nagy
egy zongora" "so big a piano" and "tu'l nagy egy zongora" "too big a
piano", is ungrammatical. Nonetheless, in exclamations, the deictic
adjective "ekkora" 'this big', when used disapprovingly, can occur in
either order: "Egy ekkora diszno'sa'g!" "a such mess" or "Ekkora egy
diszno'sa'g!" "such a mess", both meaning 'What a mess!'. Furthermore, it
is not only that Hungarian does not treat the equivalents of English
exceptional adjectival phrases as positional exceptions: their specific
selectional properties also do not show up in Hungarian. Thus, unlike in
English, "ilyen nagy" 'so big' can cooccur with plural nouns and with the
definite article; e.g.

    - ilyen nagy asztalok  "such big tables"
                           'tables so big'
    - az ilyen nagy asztal "the such big table"
                           'the table this big'

   (b) THE POST-NOUN POSITION
   No dialectal variation has been evidenced in the responses regarding
the post-noun position of exceptional adjectives (e.g. "a book this
interesting"). Thus, this order may be universal across English dialects.
Whether in other languages where the equivalents of English exceptional
adjectives are not exceptional in their prenominal ordering (i.e., they
take any article and noun and they occur after the article rather than, as
in English, before it) they are nonetheless exceptional in that they can
also occur after the noun while other adjectival phrases cannot, remains
to be seen. In Hungarian, the equivalents of English adjectival phrases
are entirely regular not only in pre-nominal position, occurring after,
rather than before, the article, but also in that, just as other
adjectival phrases, they cannot occur after the noun. Thus, the
equivalents of English exceptional adjectives are not exceptional in
Hungarian in any way, be it selection or ordering before or after the
noun. Whether in German, noun phrases such as "ein Klavier so gross" and
"ein Klavier zu gross" are grammatical or not remains to be checked.

2. ANALYSES
   (A) THE PRE-ARTICLE POSITION
   Two analyses have been proposed, one by Bingfu Lu and
one in Arnold Zwicky's 1995 article which he referred me to in his
response.

   (a) LU'S ACCOUNT
    Bingfu Lu suggests that both "this Adj" and "too Adj"
occur in front of the article because in noun phrases, modifiers
contributing more referentiality to their mother NP tend to precede
modifiers that contribute less. As supporting evidence, he cites examples
of English adjectives which precede the numeral rather than, as usual for
adjectives, follow them: "the next three lanes", "the longest three
lanes", "the left three lanes",
"the following three lanes" (cf. *"the three next lanes",
*"the three following lanes" (although "the three longest lanes" and "the
three left lanes" are, for some reason, OK)). Thus, he suggests, if we
assume that the indefinite article "a(n)" patterns with numerals, the
pre-numeral position of "this/too Adj" is covered by the same
generalization as the pre-numeral position of "next", "longest" etc.
because "this/too..." also contribute to the referentiality or
definiteness of the noun phrase. The principle or referential precedence
is laid out in more detail in Lu 1998, especially sections 4.2, 4.5 and
6.5 (e.g. pages 85-86, 99, 111, 104-105, and 173-174 in the hard copy).

   My concern with Lu's explanation is that I do not see clearly in what
way the modifiers "this/too/so Adj"
etc. contribute to making the noun phrase more referential or more
definite. The fact that "this/so/too... Adj" occurs only with indefinite
nouns militates against this idea. A phrase like "this student" is
referential and definite; but "this bright a student" is neither definite
nor necessarily referential.

   (b) ZWICKY'S ACCOUNT
   Zwicky's highly interesting paper is concerned with the general issue
of selectional relations in syntax and in particular, with constraints on
the proximity and hierarchic relation that can hold between selector and
selectee. The "this/too... Adj" construction is discussed on account of
two oddnesses. The first thing that is strange about it is that, while
"this big" is like a normal adjectival phrase in its internal composition,
its external syntax - both selection and order - is different from that of
other adjectival phrases, along the lines discussed above. Second, such
phrases select a particular determiner ("a(n)") in the NP they combine
with even though the determiner is not a sister constituent - which is the
normal relation between selector and selectee - but a "niece".

   The point of the paper is to show that, while "this/too... Adj" phrases
are exceptional, their patterning nonetheless has parallels in other
phenomena both in English and in other languages (122). For example,
the indefinite article of the noun phrase is also selected by an "aunt" in
"What a good day!" (117). The pre-article "this/too... Adj" construction
thus emerges from Zwicky's discussion as exceptional and unexplained in
its detail but not unparalleled in type.

   (B) THE POST-NOUN POSITION
   Three relevant ideas have been offered.

   (a) OTHER POST-NOMINAL ADJECTIVAL PHRASES IN ENGLISH
    Both Zwicky and Hudson noted that, just as the pre-article position is
not unique to exceptional adjectival phrases (there is also "what a day",
"such a day", etc.), the post-noun position is also not unique to them.

    Zwicky (l995) proposes that a necessary - but not sufficient -
condition for such postposable adjectival phrases is that they should also
be capable of standing as predicates (thus "late" or "former" do not
qualify). Within the bounds of this constraint, postposability is
OBLIGATORY
    - if the adjectival phrase contains a complement (e.g., "many faces
bright with joy"; cf. *"bright with joy many faces"); and
    - if the head is an indefinite pronoun (e.g. "something useful"; cf.
?"useful something"),
and it is POSSIBLE with "this/(too?)... Adj"-type adjectival phrases
(121-122). A sufficient condition for postposability - i.e. a possible
common denominator of the constructions just mentioned, is yet to be
identified.

   Richard Hudson cites a subtype of complemented adjectival phrases which
are postposable: measure expressions such as "two inches long"; cf. the
equal grammaticality of "a two-inches-long line" and "a line two inches
long" (although note the "compactness" of the preposed adjectival modifier
in comparison to the post-posed one). Since both "two inches long" and
"this big" include a noun modifier: "two inches" and "this", Hudson
suggests that this may be a common denominator between the two equally
postposable adjectival phrases and a sufficient condition for
postposability.

   This account would explain a fact noted above: that "this Adj" and
"that Adj" are postposable to the noun but "too Adj" is not or at least
not always. The explanation would be that "this" is a noun but "too" is
not. However, the problem is that this explanation is its lack of
generality: it would account for the postposability of only some
exceptional adjectival phrases: for example, "so Adj" is also postposable
just as "this Adj" ("A piano this/so big cannot fit into this room.") even
though "so" is not a noun and thus not covered by the generalization.

   However, a related point comes to mind. Whether the fact that the noun
"this" occurs as a modifier in one of the exceptional adjectival phrases
is relevant or not, remains to be seen; but note, however, that the entire
exceptional adjective phrase has something of a nominal character. This is
borne out by two facts: its peripheral - pre-article and apposition-like -
position within the noun phrase; and the option of cooccurring with "of",
a preposition which primarily occurs after nouns.

   (b) ELLIPSIS
   Several responents - Darnell, DuBartell, Lu, and Peter - related the
fact that "this Adj" can occur after the noun to the fact that such
constructions are synonymous with the corresponding relative clause
construction - "a book this good" with "a book that is this good"; and
they were considering a derivation of the former from the latter. However,
as Darnell points out, the fact that "N this Adj" is paraphrasable with a
noun plus a relative clause whose predicate is "this Adj" is not a
distinguishing feature of the "N this Adj" construction: many other
adjectival phrases are paraphrasable the same way; yet, many of those
cannot occur post-nominally. Compare "the big table", "the table that is
big", *"the table big".

3. CONCLUSIONS
   What emerges from the above regarding English exceptional adjectival
phrases is the following.

   (a) SELECTION
       - Phrase-internal selection:
         While the adjective of an exceptional adjectival phrase is freely
selectable, the modifier must come from a small set including "this",
"too", etc.
       - Phrase-external selection:
         Exceptional adjectival phrases form a proper subclass of
adjectival phrases from the point of view of their external syntax. They
are selectionally and positionally like many other adjectives in that they
can be used and ordered predicatively and in "though-fronting"
constructions just as other adjectives; but they differ from the others in
that, when preposed to the article, they cooccur only with singular
indefinite countable nouns.

   (b) LINEAR CONSTRAINTS
    Exceptional adjectival phrases must not be in the regular adjectival
post-article position in at least some dialects: instead, they are ordered
peripherally to the noun phrase. In particular, all of them must occur in
pre-article position if the precede the noun and some of them also have an
alternative post-nominal position.

   (c) VARIATION
   English dialects differ in whether they do or do not treat exceptional
adjectives as linearly exceptional when they occur before the noun - i.e.,
whether they need to be in pre-article position or not; and languages
differ the same way as well. In perhaps all English dialects, however,
exceptional adjectival phrases in pre-article position have an alternative
expression: in addition to AdjP ART N, there is also AdjP of ART N.

   This account raises a number of why-questions.

   l. SELECTIONAL CONSTRAINTS
      (a) Exactly why do just those modifiers -  "this", "too" and the
others - figure in selectionally and positionally exceptional adjectival
phrases, and not others? If referentiality, definiteness, or being
noun-like have anything to do with this, the exact conditions are yet to
be spelled out.
      (b) Exactly why are the selectional constraints of exceptional
adjectival phrases relative to their context the way they are? In
particular, why are exceptional adjectival phrases LIKE other adjectival
phrases in some of their contexts (occurrence in predicate position and in
"though-fronting") but UNLIKE them in that in prenominal position, they
cooccur with indefinite singular count nouns only?

   2. LINEAR CONSTRAINTS
      Why do exceptional adjectival phrases require or allow for just the
positions that they do; and why do they differ among each other in whether
they can be post-nominal or not?

   3. VARIATION
      - Why is there both an "of-less" and an "of-full" variety when
exceptional adjectival phrases stand before the article?
      - Why is it that these adjectival phrases are exceptional in some
dialects of English but less exceptional in other dialects?
      - Why are the equivalents of English exceptional articles not
exceptional in some other languages?

   Regarding this last issue, there are four possibilities. First,
exceptional adjectival phrases may be a random characteristic of English,
similar to any lexical feature such as that 'table' is "table" and not
"stol". Second, their presence may be predictable from some other property
of English grammar, just as the presence of /k/ is predictable from the
fact that English has a /p/ if a universal implication holds according to
which languages with /p/ always have a /k/. Or the presence of exceptional
adjectives may be it like the /p/, so that it predicts some other
characteristic of English grammar. And, fourthly, this feature may be in a
bidirectional implicational relationship with some other feature both
predicted by it and predicting it. As things now stand, the first, least
desirable option holds, according to which the presence and particular
properties of exceptional adjectival phrases are an isolated and thus
accidental - rather than a systemic, or integrated - property of English.

REFERENCES AND ADDITIONAL BIBLIOGRAPHY

The following relevant titles have been mentioned in the responses. I have
not been able to consult Abney 1987 (324), Baker 1989 (327), and Hudson
1990 (370-371).

Abney, Steven P. 1987. _The English noun phrase in its
   sentential aspect._ Ph.D. dissertation, MIT.

Baker, C.L. 1989. _English syntax._ Cambridge, MA: MIT
   Press.

Hudson, Richard. 1990. _English word grammar._

Lu, Bingfu. 1998. _Left-right assymmetries of word order
   variation: a functional explanation_ Ph.D. dissertation,
   University of Southern California. Available at
http://www.usc.edu/dept/LAS/ealc/chinling/articles/lup.pdf

Radford, Andrew. 1993. "Head-hunting: on the trail of the
   nominal Janus" In  Greville G. Corbett, Norman M.
   Fraser, and Scott McGlashan, ed., _Heads in grammatical
   theory_. Cambridge; Cambridge University Press. 73-113.

Zwicky, Arnold. 1995. "Exceptional degree markers: A puzzle
   in internal and external syntax" _Ohio State University
   Working Papers in Linguistics_ 47, 111-123.



   ************************************************************************
                         Edith A. Moravcsik
                         Department of Foreign Languages and Linguistics
                         University of Wisconsin-Milwaukee
                         Milwaukee, WI 53201-0413
                         USA

                         E-mail: edith at uwm.edu
                         Telephone: (414) 229-6794 /office/
                                    (414) 332-0141 /home/
                         Fax: (414) 229-2741