LINGUIST List 7.27

Sat Jan 6 1996

Disc: Levin & Rappaport Hovav (1995)

Editor for this issue: Ann Dizdar <>


  1. Richard Ingham, Re: 7.2, Disc: Levin & Rappaport Hovav (1995)

Message 1: Re: 7.2, Disc: Levin & Rappaport Hovav (1995)

Date: Thu, 04 Jan 1996 15:07:42 GMT
From: Richard Ingham <>
Subject: Re: 7.2, Disc: Levin & Rappaport Hovav (1995)

As regards unaccusativity, the terminology issue is perhaps less
important than an understanding of what the unaccusative hypothesis
itself has to offer. In this connection, I wonder to what extent
participants in this LINGUIST list discussion share my impressions of
Levin & Rappaport-Hovav (1995).

What I found striking in the book was the extent to which the work of
argument selection in English was being done by membership of various
semantic subclasses, NOT by the argument-structure property of whether
the argument is internal or external. I don't see this as a criticism,
except insofar as the authors maintain that unaccusativity is
'syntactically encoded' in the form of a lexical syntactic structure
that distinguishes internal from external arguments (see e.g. pp. 279,
For example, the causative alternation is impossible with many verbs
that are supposed to have only an internal argument, e.g.:

1) They ... *arrived the goods/*fell the lamp/*departed the guests/*receded 
the danger

2) They ... *languished the prisoners in jail/*persisted the recession/
*prospered the economy

But a form of the alternation is found with verbs that L&R (p.81) 
describe as 'unergative':

3) Mary walked the dog in the park/galloped the horse on the main road

The availability of an argument structure with an internal but no
external argument (the crux of the GB unaccusativity hypothesis) is
thus not decisive for the syntactic alternation between Subject and
Object NP found with certain English verbs. To handle cases like
1)-3), L&R bring in various additional assumptions, involving semantic
notions such as 'verbs of appearance', 'externally caused' and
'coercion'. Again, I don't object, except that the work is being done
by these semantic notions, not by argument alignment.

To my mind, the least convincing part of the book is where L&R find
that diagnostics for unaccusative/unergative class membership clash,
as with agentive verbs of manner of motion. They then take the ad hoc
step of postulating two lexical entries. The unaccusative one seems to
be needed ONLY to deal with the presence in the sentence of words from
what L&R (p.186) describe as a restricted number of resultative
predicates such as
 _clear_, _free_ and _apart_. Independent evidence for the unaccusative 
entry is lacking:

4) *The previously jumped/swum contestant
5) *There jumped/swam a prisoner (free)
6) *a jumpee, *a swimmee	

An unaccusative lexical representation with these verbs may get the
theory out of a jam, but doesn't provide any particular insight. On
the contrary, postulating that a monadic verb needs to specify its
argument syntactically as internal or external (the unaccusative
hypothesis) is what causes the problem here, it seems to me.

After all this negative stuff, let me say wholeheartedly that I found
the book an absolute delight to read, with a wealth of insightful
discussion of the syntax and semantics of verbs in English and
numerous other languages. But, at least as regards English, the idea
that some surface intransitive verbs have an internal argument in
their lexical syntactic representation doesn't convince this
particular reader: a semantic account of 'unaccusative' phenomena
(e.g. Dowty 1991) still gets my vote.

Richard Ingham
Department of Linguistic Science
The University of Reading
Reading RG6 6AA
Mail to author|Respond to list|Read more issues|LINGUIST home page|Top of issue