Who Posh Have I Been Talking to? An Investigation into a Class of Non- Standard Adjectival Wh-phrases in English.
- Timothy O. Bazalgette, University of Cambridge
- Old College Lorimer Room, University of Edinburgh
In many dialects of English there exists a class of related wh-constructions, characterised by a
wh-element followed by an adjective and no modified noun phrase. Examples of these
constructions are shown below, with the relevant wh-phrase indicated with square brackets:
(1) [Who posh] have I been talking to?
(2) [Where nice] would you like to go?
(3) [What big] happened yesterday?
These constructions (henceforth all referred to with the blanket term who-posh constructions)
occur almost exclusively in spoken form, and tend to be expressed in specific discourse contexts
and with a specific intonation profile, and so seem somewhat ungrammatical when written and in
isolation. They are nonetheless attested in native speakers, as in the following (spoken) example
by a speaker from the south-west of the UK taken from the British National Corpus (KBL 4515):
(4) [What different] have you got ta do a, a tape a day or something?
In my presentation, I intend to make it clear that who-posh constructions have various
interesting and unexpected properties, and show that attempts to account for these properties may
reveal interesting facts about the nature of English wh-phrases in general.
I first present some preliminary evidence for the geographical distribution of these
constructions, which demonstrates that the types of construction used varies largely throughout
English-speaking countries. This variation, however, is not entirely random, with only certain
combinations of constructions being used, suggesting an implicational hierarchy which makes clear
suggestions about the emergence of the phenomenon in apparent time (Labov 1966). I then
investigate the licensing of the construction in some detail, considering both the wh-elements
which may enter into who-posh constructions as well as the constraints on their acceptability. In
general, the constructions are more acceptable when pragmatically salient (such as when the
adjective involved is prominent in the discourse), when the adjective is phonologically light, and
when it is semantically unmarked. All these effects, including the phonological ones, appear to be
constraints on performance rather than on competence.
Having reviewed the distribution and constraints on usage, I then move on to a formal analysis,
first considering the semantics of the constructions, where it becomes clear that the lack of a noun
alongside adjectival modification is somewhat problematic. This difficulty is also reflected in the
syntax, and here a number of other problems arise – for instance, the construction is limited to whphrases
and has no declarative equivalent, and also behaves variably with respect to D-linking
(Pesetsky 1987). Taking a minimalist (Chomsky 1995) approach, I assess two putative analyses of
the construction – one involving a covert light noun within the wh-phrase, and another making use
of a reduced relative clause. I conclude that the latter of these is the most productive approach, as
it accounts for more of the difficulties as well as, for instance, linking the construction to Tsai’s
(1994) decomposition of English wh-phrases. Such a conclusion not only explains who-posh
constructions, but also sheds light on the internal structure of wh-elements in English.
wh-element followed by an adjective and no modified noun phrase. Examples of these
constructions are shown below, with the relevant wh-phrase indicated with square brackets:
(1) [Who posh] have I been talking to?
(2) [Where nice] would you like to go?
(3) [What big] happened yesterday?
These constructions (henceforth all referred to with the blanket term who-posh constructions)
occur almost exclusively in spoken form, and tend to be expressed in specific discourse contexts
and with a specific intonation profile, and so seem somewhat ungrammatical when written and in
isolation. They are nonetheless attested in native speakers, as in the following (spoken) example
by a speaker from the south-west of the UK taken from the British National Corpus (KBL 4515):
(4) [What different] have you got ta do a, a tape a day or something?
In my presentation, I intend to make it clear that who-posh constructions have various
interesting and unexpected properties, and show that attempts to account for these properties may
reveal interesting facts about the nature of English wh-phrases in general.
I first present some preliminary evidence for the geographical distribution of these
constructions, which demonstrates that the types of construction used varies largely throughout
English-speaking countries. This variation, however, is not entirely random, with only certain
combinations of constructions being used, suggesting an implicational hierarchy which makes clear
suggestions about the emergence of the phenomenon in apparent time (Labov 1966). I then
investigate the licensing of the construction in some detail, considering both the wh-elements
which may enter into who-posh constructions as well as the constraints on their acceptability. In
general, the constructions are more acceptable when pragmatically salient (such as when the
adjective involved is prominent in the discourse), when the adjective is phonologically light, and
when it is semantically unmarked. All these effects, including the phonological ones, appear to be
constraints on performance rather than on competence.
Having reviewed the distribution and constraints on usage, I then move on to a formal analysis,
first considering the semantics of the constructions, where it becomes clear that the lack of a noun
alongside adjectival modification is somewhat problematic. This difficulty is also reflected in the
syntax, and here a number of other problems arise – for instance, the construction is limited to whphrases
and has no declarative equivalent, and also behaves variably with respect to D-linking
(Pesetsky 1987). Taking a minimalist (Chomsky 1995) approach, I assess two putative analyses of
the construction – one involving a covert light noun within the wh-phrase, and another making use
of a reduced relative clause. I conclude that the latter of these is the most productive approach, as
it accounts for more of the difficulties as well as, for instance, linking the construction to Tsai’s
(1994) decomposition of English wh-phrases. Such a conclusion not only explains who-posh
constructions, but also sheds light on the internal structure of wh-elements in English.