|AUTHOR: D'Alessandro, Roberta.
TITLE: Impersonal Si-Constructions
SUBTITLE: Agreement and Interpretation
SERIES: Studies in Generative Grammar 90
PUBLISHER: Mouton de Gruyter
Jan Schroten, Department of Foreign Languages, Utrecht University
This research monograph investigates the properties of Italian impersonal
si-constructions. In impersonal si-constructions, the 3rd person reflexive
clitic pronoun si has no reflexive interpretation; it is used in constructions
in which an impersonal, arbitrary subject is understood. Impersonal
si-constructions, abbreviated as ISCs by the author, have a number of peculiar
properties which are described and discussed in the framework of ''minimalist
In the first, introductory chapter, the author presents an overview of
generative studies on impersonal si. Analyses of impersonal si constructions, or
ISCs, are introduced and compared. Properties that were discovered and analyzed
and the different solutions that have been proposed for some problems in the
context of generative grammar are evaluated in much detail throughout this
monograph. Strong and weak points of theoretical approaches based on different
descriptions of ISCs and different generative models are taken into
consideration. The version of the ''minimalist program'' that she has adopted as
the basis of her treatment is introduced, with special attention to the parts
that turn out to be useful to provide an explanation for the peculiar phenomena
found in ICSs.
Four phenomena are chosen as special problems, and chapters two to five present
these phenomena and try to give a ''minimalist'' explanation. The four phenomena are:
(i) Agreement vs. lack of agreement of the tensed verb with the ''underlying''
direct object in constructions that are called ''passive si-constructions'' (with
agreement) and ''transitive si-constructions'' (without agreement).
(1) In Italia si mangiano<3pl> gli spahetti [with agreement]
(2) In Italia si mangia <3sg> (gli) spaghetti [without agreement]
(In Italy they eat spaghetti)
(ii) Person restrictions in transitive ISCs.
(3) In televisione si vede<3sg> spesso lui<3sg.Nom> / si vedono<3pl> spesso
(One often sees him / them on the tv)
(4) *In televisione si vedo<1sg> spesso io<1sg.Nom> / si vediamo<1pl> spesso
(One sees often me / us on the tv)
(iii) The ''inclusive interpretation'' of impersonal si, that is, the reasons why
the pragmatically arbitrary or impersonal referent is, or is not, inclusive of
(5) In quel ristorante si mangiava bene [noninclusive interpretation]
(One used to eat well in that restaurant)
(6) In quel ristorante si è mangiato bene [inclusive interpretation]
(Somebody/we have eaten well in that restaurant)
(iv) Past participle and adjective agreement and non-agreement in ISCs.
(7) Si è telefonato
(They/we have called)
(8) Si è arrivati / arrivate
(They/we have arrived)
The phenomenon of agreement vs. lack of agreement (cf. (i) above) has always
been discussed in treatments of ISCs. In the grammatical tradition, lack of
agreement has been qualified as substandard or simply wrong. Less prescriptive
studies observed that lack of agreement was frequently found, with a slightly
The question has always been what kind of difference there is. The understood
arbitrary subject of (2) has no different interpretation than the arbitrary
subject of (1). What was observed is that indefiniteness of the underlying
direct object seemed to favor lack of agreement. The difficult point was that no
explanation could be given for this apparent correlation.
D'Alessandro hypothesizes that agreement and lack of agreement correlate with
verbal semantics or Aktionsart: activities favor lack of agreement and
accomplishments favor agreement.
In her interpretation of minimalist theory, lack of agreement in ''transitive''
ISCs correlates with treating si as a subject pronoun, attached to v-VP, whereas
agreement correlates with ''accusative'' si, forcing the underlying direct object
to behave as a subject.
The phenomenon of person restriction in transitive ISCs (cf. (ii)) is described
in great detail, and compared with seemingly related, but different properties
of Icelandic quirky restrictions.
The general picture that arises is that impersonal si takes part in ''multiple
agreement'' with the underlying direct object. Since si has <3rd person>
specification, the agreeing direct object must match under multiple agreement,
and be <3rd person> too; otherwise, multiple agreement is blocked, due to a
mismatch of <3rd person> si with, for example, <1st person> io, noi, as in (4).
D'Alessandro addresses inclusive interpretation of impersonal si constructions
(cf. (iii)). Basically, ''impersonal si constructions'', or ISCs, have generic
and/or existential interpretation; inclusive interpretation is obtained when the
speaker is part of the individuals with respect to the existential
interpretation. Of course, specific time reference favors existential and
inclusive interpretation, whereas lack of temporal restrictions favors generic
interpretation. The examples given in (5) and (6) illustrate this well-known fact.
In this monograph, considerable attention is given to the question how the lack
of ''event boundaries'', which triggers generic interpretation, is expressed and
how the ''event boundaries'' are expressed triggering existential and inclusive
The oddest phenomenon found in ISCs are the plural past participles of
unaccusative or ergative verbs and of adjectives, which can take the feminine
plural form if the generic or existential referent is feminine (cf. (iv)). Past
participles of other verbs take the default, masc. sg., form.
D'Alessandro argues that the number feature is somehow sub-divided in a number
and animacy feature, including the feature, traditionally called arb (arbitrary
subject). She suggests a possible interpretation of these facts in the
D'Alessandro's research monograph reflects the long history in generative
grammar of studies on impersonal si. The extensive presentation and discussion
of studies of impersonal si is useful and necessary. The selection of the four
phenomena is adequate, in some cases traditional, in other cases novel and
The approach to the first and second phenomenon is very good and merits much
attention due to the original and novel ways of presenting vague but consistent
intuitions and accounting for them in the minimalist framework.
At the same time, it must be said that the presentation and discussion of
studies on impersonal si reflects the Ph. D. dissertation origin. The reader
would have benefited from a shorter, and more selective, presentation and
discussion, focusing the issues that are crucial for the proposals that the
author wishes to make.
The first two phenomena, agreement and lack of agreement in si-constructions of
transitive verbs and the person restrictions are handled with much care and
benefit from the discussion and the observations that are made.
The third and fourth phenomena, inclusive interpretation and past participle
agreement in impersonal si-constructions, are interesting, but D'Alessandro's
proposals are less convincing: they require a very limited model of minimalist
theory. In many cases, the author stresses the fact that convincing further data
are unavailable and that other approaches can be taken.
The weakest point of this monograph is that no hint or suggestion is given as to
why impersonality is so much linked to (superficial) subjects and why impersonal
objects are expressed in very different ways, if they can be expressed at all.
In short, this monograph deserves many readers. Readers will be surprised by the
original proposals and will be stimulated to find solutions for pending problems
ABOUT THE REVIEWER
Jan Schroten Ph.D. is a - recently emeritus - associate professor of Spanish
linguistics at Utrecht University. His research projects are on syntax,
morphology and the lexicon of Spanish: clitic doubling as a kind of agreement,
and: meaningful suffixes in the lexicon of Spanish and how to account for them.