Mateus, Maria Helena and Ernesto d'Andrade. 2000. The
phonology of Portuguese. Oxford: Oxford University Press.
Pp. x, 162. HB. $74.
Andrew J. Koontz-Garboden, Indiana University at Bloomington
The book under review appears as part of the Oxford
University Press series on the phonology of the world's
languages, whose purpose is to offer a theoretical overview
of the phonology of one language. In the case of the
present volume, the language is Portuguese, and it is
described within the framework of rule-based autosegmental
Lexical Phonology. Although both continental and Brazilian
varieties are examined throughout, the basis for discussion
is the continental variety spoken in Lisbon and Coimbra (4).
In what follows, a synopsis and critical evaluation of the
book in question are given.
Chapter one deals with such preliminary issues as
orthography, transcription system, and the history of
Portuguese. Likewise, the first half of chapter two is
largely introductory, giving a nontheoretical introduction
to Portuguese phonology (10-23). It is in the second half
of chapter two that M&A lay out some of their theoretical
assumptions. They assume a constriction-based model of
feature organization based upon Clements and Hume (1995)
with some minor modifications that in most cases appear to
be specific to Portuguese (e.g. they (28) reject scalar
vowel height values argued by others to be necessary for
other languages). Following discussion of the feature
geometry system, M&A give (fully specified) feature charts
for the consonant, vowel, and glide systems of Portuguese
(29-30). Next, they walk the reader through the basic
reasoning behind their (radically) underspecified account of
Portuguese vowels and consonants (31-33). With respect to
vowels, they find that /i/ is unmarked, and accordingly the
value [+high] is to be filled in (along with redundant
features) during the course of the derivation. In the
consonantal realm, [+anterior] coronal segments are argued
to be unmarked, and accordingly the values of the C-place
node for these segments are filled in during the course of
the derivation (along with redundant features). Feature
tables representing the underspecified lexical
representation of the vocalic system (35) and the
consonantal system (36) are also given.
Chapter three, which treats syllable structure, begins with
a discussion of possible onsets and codas following the
Sonority Sequencing Principle. In their analysis of onsets,
M&A argue in favor of empty surface nuclei basing themselves
upon surface alternations (in the continental variety only)
such as [de'vu] 'I owe', [dve'r] 'to owe', and [dvdo'r]
'debtor' (44-46). They adduce dialectal and child language
evidence in favor of their proposal that the nucleus vowel
in such alternations is present in the underlying
representation, but deleted at the surface in continental
Portuguese, leaving an empty nucleus. Next they discuss
glides and diphthongs by appealing to the notion of the
rhyme (46-52). Their position is that there are no lexical
glides in Portuguese; these are derived from underlying /i/
and /u/ which can become glides when syllabified as part of
the nucleus (when following another vowel) (46) or in the
onset (when preceding a vowel) (51). Following this, they
turn their attention to possible coda consonants. The
discussion centers around the fact that the only possible
coda consonants in Portuguese are /l/, /r/, and /s/. In
this section further vowel epenthesis evidence from BP is
adduced in favor of the postulation of empty nuclei.
Elaboration of the skeletal tier follows discussion of basic
syllable structure. X-slot theory (as opposed to CV theory)
is adopted, and processes relating to the presence of empty
X positions are discussed (as in e.g. the realization of a
floating nasal feature as a nasal consonant(59)).
Additionally, they posit segments without skeletal
positions, and exemplify these by distinguishing light
diphthongs (resulting from epenthesis) from heavy (lexical)
diphthongs (55-58). Finally, they discuss syllabification
rules, which, when the proper conditions hold, associate
nuclei (60), associate onsets (61), create empty nuclei
(62), create empty onsets (62), and associate codas (63).
It is not made entirely clear exactly where syllabification
first takes place. At times M&A seem to suggest that they
are taking the somewhat non-standard assumption that the UR
is enriched with syllable structure. This is the case in
e.g. their discussion of the syllabification rules where
they (64) say that ".the underlying syllables differ
crucially from those on the phonetic level.." It is
possible that what is meant by "underlying syllable" in this
context is the resulting form after syllabification on the
first cycle, although this is not stated. The same sort of
ambiguity arises in their discussion of diphthongs (48-51)
(e.g. "We propose that these phonetic glides are vowels
underlyingly, and they are nuclei of independent syllables."
Chapters four and five deal with morphophonemic processes
assuming a rule-based model of Lexical Phonology and
Morphology (Kiparsky 1982). In chapter four M&A concentrate
on inflectional processes such as gender and number
inflection in nouns and adjectives and verbal inflection.
Two of the more important phenomena dealt with under these
headings are the phonology of number inflection and vowel
harmony in verbs. The various allomorphs of the Portuguese
plural morpheme /s/ are argued (70-73) to be the result of a
series of morphology-sensitive phonological rules (i.e.
phonological rules taking places in the lexical phonology)
that are for the most part sensitive to syllable structure.
Following Wetzels (1991), vowel height harmony is treated as
autosegmental spreading of the height features of the theme
vowel onto the stressed vowel, following theme vowel
deletion (82-83). In addition to this, there is claimed to
be an independent lowering rule, which is presumably
crucially ordered after the vowel harmony rule, although
this is not entirely clear. Although mentioned only in
passing, it seems crucial to their account that vowel
harmony is a lexical phenomenon, since there exist
exceptions, as they note (85). Chapter five is a discussion
of derivational morphology again from a rule-based lexical
phonology perspective. Such theoretical notions as roots,
stems, and words are introduced and exemplified with
Portuguese lexical items, while various idiosyncratic
properties of different Portuguese derivational affixes are
Chapter six treats word-level stress. M&A develop a
rule-based analysis whereby Portuguese has two separate
systems for assigning primary stress: one system for verbs,
and a separate system for nouns and adjectives. Lexical
items can also be marked in the lexicon for exceptional
stress patterns. In these cases the rules do not apply.
Given this set of circumstances, it seems clear that they
are assuming stress assignment to take place in the lexical
phonology, although this is never made explicit. In the
subsequent grid-based analysis that they develop, the entire
system is basically analyzed as trochaic, with right-to-left
footing, and a right-headed metrical word. For nouns having
no class marker, M&A posit a null class marker that has a
rhythmic position, and is still used in the calculation of
stress assignment (123). In this way, stress on such words
(e.g. cafe', hospita'l) appears as regular. Additionally,
they assume two main forms for lexical marking of stress:
certain morphemes are "stress repellent" (124), and words
with generally idiosyncratic stress, or antepenultimate
stress, if not accounted for by a stress repellent morpheme,
can be lexically marked (124).
Chapter seven treats several different phonological process
not already discussed earlier. Nasal vowels, as is hinted
at in previous chapters, are treated as the realization of a
floating nasal autosegment (130). The floating autosegment
is realized as a nasal consonant when it precedes an empty
onset, but nasalizes a vowel when there is no empty onset
position to fill. This process crucially applies in the
lexical phonology and accounts for several different
morphological alternations, such as e.g. the two different
realizations of the prefix /in-/ 'not,' as in (1) and (2)
(1) incapaz [i~] 'unable
(2) inacabado [in] 'unfinished'
In (1) the nasal autosegment can not fill the onset
position, since this is already filled by a consonant. It
therefore nasalizes the vowel. In (2), however, the onset
is empty, and can be filled by the feature, which is then
realized as the unmarked alveolar nasal. Also discussed in
this chapter are continental Portuguese vowel reduction
(134-136), processes conditioned by syllable structure
(137-144) (e.g. Brazilian Portuguese /l/-->[w] in coda
position), and some connected speech (i.e. sandhi) phenomena
COMMENTARY AND CRITICISM
_The phonology of Portuguese_ succeeds in introducing the
reader to some important problems in Portuguese phonology.
It does, however, have some shortcomings, which I discuss
One problem with the work under review is the fact that it
is extraordinarily short (only 162 pages including
references, author index, and a short subject index; only
148 pages of text). Indeed, it is half the size of most
other books in the same series (e.g. Odden 1996, Hammond
1999, Weise 2000). The result of this is at times a lack of
explicitness and development in the discussion, as might be
expected from such a short book with such a broad objective.
An example of a problem related to this is the discussion of
the relationship between the vowel harmony rule and the
vowel lowering rule (86). In rejecting Harris' (1974)
analysis, M&A claim (86) that they ".accept Wetzels'
arguments maintaining that L[owering] R[ule] interacts with
the morphological rules and it applies at the lexical
level..This proposal allows the LR to be a filling-rule that
fills the feature values of the root vowel that is not
lexically specified." Unfortunately, the discussion is left
at this; neither do they clearly spell out what Wetzels'
(1991) arguments are against Harris' elsewhere condition
analysis, nor do they formalize the filling-rule. The
former problem is particularly regrettable in a volume of
this nature, since Wetzels (1991) is an article written in
Portuguese, and one of the objectives of the authors (as
stated in their preface) is to make available and expand
upon analyses that have thus far appeared only in
Portuguese. The latter, problem (the lack of formalization
of the filling-rule) is also bothersome since this rule
potentially interacts with the vowel harmony rule.
In other areas of the book, M&A appear to take some
non-standard assumptions without really justifying them as
well as would be expected. This is the case with the
floating nasal autosegment, which appears to be linearly
ordered on the segmental tier, crucially at the place where
it needs to be in order to cause vowel nasalization (131).
Likewise, in the discussion of stress assignment M&A (123)
assume a phonetically null class marker with a rhythmic
position without theoretical justification and without
discussion of the implications for other processes.
An additional criticism, is that the book includes neither
discussion of rule ordering, nor even a list of all of the
rules that have been posited (although a number of these are
listed with page references in the subject index).
Likewise, it is not always clear exactly when (i.e. at what
point in the derivation and in what order with respect to
other rules) features are filled in as part of the
underspecification account. Additionally, no full
derivations are given to demonstrate their analyses. The
lack of discussion on rule-ordering and interaction makes it
difficult to globally evaluate their analysis. The lack of
full derivations for the analyses of particular problems
makes it at times difficult to see how one rule interacts
with another (e.g. vowel harmony and vowel lowering
The theoretical approach to the analysis in the book should
also be highlighted; as mentioned above, it is carried out
within a serial, rule-based, Lexical Phonology approach (it
should be pointed out that they do not argue in favor of
this theory, they merely adopt it as a descriptive device).
That they prefer such an approach is, of course, not
objectionable in and of itself. What is bothersome,
however, is the lack of comparative perspective, especially
in light of the increasing acceptance of Optimality Theory
(Prince and Smolensky 1993) among phonologists. In only one
instance is an OT analysis even mentioned (this is M&A's
brief discussion (73) of Morales-Front and Holt's (1997) OT
analysis of Portuguese pluralization). This may be the
result of the short length of the book. It is,
nevertheless, unfortunate, especially for the reader who is
ultimately interested in phonological theory as opposed to
the phonology of a single language.
A final criticism of the volume under review is the
apparent absence of sufficient editorial support; typos, and
stylistic infelicities are more numerous than one would
expect from a book published by OUP. This makes at times
for difficult reading. Some examples include: the variant
spelling of an author's name (84) (Wetzels vs. Wetzel); the
use of variant possessive orthography with the same author's
name (84, 86) (Wetzels's vs. Wetzels'); reference to an
example claimed to be a rule that is, in fact, not a rule,
but rather part of a feature geometry representation ((19)
on page 78); and lack of consistency in the use of some
terminology (e.g. the penultimate syllable is varyingly
referred to as: "the syllable before the last" (109), "the
penultimate" (110), "the syllable preceding the final one"
(113), "the second from the end" (114), and "the last but
one" (117)). Additionally, one finds some lexical and
syntactic peculiarities such as several bizarre uses of the
suffix -ly (e.g. ".as it may be assigned to a previously
trough position." (125)).
Despite these criticisms, _The phonology of Portuguese_
will be of use to those who want an introduction to the basic
issues in Portuguese phonology, especially for those who do
not read Portuguese. Likewise, it will serve as a useful
reference since it is fairly data-rich (given its length),
and since it brings together in one place analysis and
discussion (although brief) of the major phonological
phenomena of the language.
Clements, George N. and Elizabeth Hume. 1995. Internal
organization of speech sounds. In J. Goldsmith (ed.). _The
handbook of phonological theory_. Oxford: Blackwell. Pp.
Hammond, Michael. 1999. _The phonology of English: a
prosodic optimality-theoretic approach_. New York: OUP.
Harris, James. 1974. Evidence from Portuguese for the
elsewhere condition. _Linguistic inquiry_ 5: 61-80.
Kiparsky, Paul. 1982. Lexical phonology and morphology.
In I.-S. Yang (ed.). _Linguistics in the morning calm_.
Seoul: Hanshin. Pp. 3-91.
Morales-Front, Alfonso and D. Eric Holt. 1997. On the
interplay of morphology, prosody, and faithfulness in
Portuguese pluralization. In Fernando Martinez-Gil and
Alfonso Morales-Front (eds.). _Issues in the phonology and
morphology of the major Iberian languages_. Washington,
D.C.: Georgetown University Press. Pp. 393-437
Odden, David. 1996. _The Phonology and Morphology of
Kimatuumbi_. New York: OUP.
Prince, Alan and Paul Smolensky. 1993. Optimality theory:
constraint interaction in generative grammar. RuCCS
Technical Report 2. Piscataway, NJ: RuCCS.
Wetzels, W. Leo. 1991. Harmonizacao vocalica, truncamento,
abaixamento e neutralizacao no sistema verbal do portugues:
um analise auto-segmental. _Cadernos de estudos
linguistics_ 21. Campinas: UNICAMP-IEL.
Wiese, Richard. 2000. _The phonology of German_. New
Andrew J. Koontz-Garboden is in the Department of
Linguistics and the Department of Spanish and Portuguese at
Indiana University at Bloomington. His interests are in
phonological and syntactic theory. Among other things, his
recent work in these areas has focused on the formal
analysis of phonological opacity in optimality theory, and
on syntactic markedness in pidgin and creole languages.
Andrew J. Koontz-Garboden
Department of Linguistics and
Department of Spanish and Portuguese
Indiana University / BH848
Bloomington, IN 47405 U.S.A.