LINGUIST List 15.560

Thu Feb 12 2004

Disc: Re: "double be" and other non-standard BEs

Editor for this issue: Sarah Murray <>


  1. Patrick McConvell, Disc: English double-be and other be's
  2. Sebastian Ross-Hagebaum, RE: "double be" and other non-standard BEs

Message 1: Disc: English double-be and other be's

Date: Thu, 12 Feb 2004 09:28:03 +1100
From: Patrick McConvell <>
Subject: Disc: English double-be and other be's

I would like to comment on Jilll Murray's suggestion (Linguist 15.535,
February 9 2004) about the distribution of what I have called the
'classic' double-be construction; and Sebastian Ross-Hagebaum's
response to my response on his comments on my summary.

Jill Murray says that classic 'double-be' is not found in Australian
English and suggests that it may be regionally restricted, pointing to
Irish English as one locus. In my original posting I stated that this
phenomenon has been found widespread in all first language
English-speaking countries (although I have not checked South
Africa). Most of the examples in my 1988 article were from Australia
and I hear it virtually every day from Australian born colleagues, on
the radio etc. Where it started is a different question--the earliest
examples I have are from US films in the early 1980's but I have not
done a systematic search of corpora or other sources.

Ross-Hagebaum says that my approach has a different purpose to his, in
being diachronic and concerned with prosody rather than syntax, but at
the same time suggests that my approach does not take into account
information structure. He also calls into question my proposed
constraint on blends or amalgams.

Perhaps I am unorthodox here but I do not believe the synchronic study
of variation can be so easily divorced from the diachronic study of
'change in progress', and we may lose valuable insights if we try to
separate the two approaches too radically. Similarly the question of
prosody is intimately bound up with both the classic 'double be' and
the other 'be' discussed by Ross-Hagebaum, which cannot be treated as
purely syntactic. Prosody is a way of conveying information packaging
alongside syntax, and 'double be' results from a disjunction of syntax
and prosody/information structure in my view. I do discuss the
information structure of 'double be' in the 1988 article as it is
notable that the apparent 'subordinate clause' is the main assertion
in information terms. In that paper I do not suggest an information
analysis of the other BE, which I might call FREE-BE. R-H proposes
that the structure is Presupposition-Focus and is thus related to the
pseudo-cleft syntactic structure of one of the blend components. My
1988 paper includes a wider range of phenomena with FREE-BE's than
those discussed by R-H, not restricted to sentences beginning 'That's,
and the analysis might require reconsideration in the light of that.

Regarding my proposal that syntactic blends should be restricted to
phenomena in which two constructions which are virtually synonymous
merge, I am aware that previous discussion of amalgams has included a
lot of constructions which do not conform to this description. I have
my doubts about some of these analyses. Given that we have two
virtually synonymous terms floating around here, I would like to claim
the term 'blends' for those conforming to the strict criteria I have
proposed, and assign 'amalgams' to the wider class.

- Patrick McConvell

Mail to author|Respond to list|Read more issues|LINGUIST home page|Top of issue

Message 2: RE: "double be" and other non-standard BEs

Date: Tue, 10 Feb 2004 22:19:01 -0500 (EST)
From: Sebastian Ross-Hagebaum <>
Subject: RE: "double be" and other non-standard BEs

I would like to respond to Patrick McConvell's post from February 9
[Linguist 15.535]. I hope to do McConvell justice in summarizing his
position as follows:

McConvell's primary concern is the origin and development of the
double-BE construction. His hypothesis is that the occurrence of the
extra BE is a process of "lexicalization of prosody": in a prosodic
context characterized by a sequence of "high tone, brief pause,
resumed high tone", BE can come in and replace this prosodic
feature. That is essentially his analysis of the "classic" double-BE
construction (e.g., The thing is, is that ...; What I mean is, is
that...; see McConvell 1988 and others; see [Linguist 15.427]). For
what I have called the "that's X is Y" construction, McConvell
suggests the same origin. Thus, according to him, a sentence like:

1) That's what I liked about her too: she fixed her own car.

is characterized by the same prosodic feature of "high tone, brief
pause, resumed high tone" (represented orthographically by the colon),
and consequently exhibits the right condition for the replacement of
the "colon-intonation" by BE. This occurs on the model of:

2) What I liked about her too is she fixed her own car.

so that 1) and 2) blend into:

3) That's what I liked about her too is she fixed her own car.

His hypothesis is that the "classic" double-BE is an earlier
phenomenon and that BE as lexicalized prosody has subsequently made
its way into "extended" contexts such as 3).

My own amalgam analysis of the sentence in 3) [Linguist 15.518] is one
that considers the two constructions in 4) and 5) as "blend

4) That's what I liked about her too.
5) What I liked about her too is she fixed her own car.

My own concern is entirely different from that of McConvell. I am
interested in "that's X is Y" as a construction present in the grammar
of English. As such, I view it as a construction in its own right, not
reducible to (derivable from) other constructions. However, it is
related to other constructions. This relationship can be captured by
appealing to the notion of amalgam or blend. The synonymy constraint
that McConvell imposes on amalgams/blends (he describes 1) and 2) as
"virtually synonymous") is, in my opinion, not required, because
amalgams/blends do not need to be traced back to two synonymous
sources (see, e.g., Lambrecht's (1998) analysis of sentences like
"There was a farmer had a dog").

An aspect of my analysis that has not been mentioned yet is that it
tries to take into account the discourse use of "that's X is Y" tokens
to motivate a characterization of the information structural
properties of the construction. In an abbreviated way, this can be
represented as follows (where FOC y"ocus and PRES y"resupposition).

6) [That's]FOC (what I liked about her too)PRES is [she fixed her own

This proposal builds on what is known about the information structural
properties of (reverse) wh-clefts and the other "blend
components". While I am *not* saying that "that's X is Y" derives
compositionally from them, it *is* related to them in that it shares
certain formal and functional characteristics.

This does seem to fall within the domain of the analysis proposed by
McConvell. And while both "classical" double-BE and "that's X is Y"
may have developed via the same mechanism of language change (as
suggested by McConvell's lexicalized prosody hypothesis), we still
need to capture the differences in form and function that presently
characterize them.

- Sebastian Ross-Hagebaum


Lambrecht, Knud. 1988. There was a farmer had a dog: Syntactic
amalgams revisited. BLS 14, 319-339.

Subject-Language: English; Code: ENG 
Mail to author|Respond to list|Read more issues|LINGUIST home page|Top of issue