ArticlePDF Available

On the partially divergent phonology of Spanish, Portuguese and points in between



In this chapter we explore the comparative markedness and phonotactics of Spanish and Portuguese phonology as a framework in which to consider: (i) the segmental inventories of each language, (ii) the phonological processes of the grammar of each language (including how prosody may impact these), as well as (iii) certain historical and dialectal developments, including (iv) phonologically intermediate/hybrid variants.
is is a contribution from Portuguese-Spanish Interfaces: Diachrony, synchrony, and contact.
Edited by Patrícia Amaral and Ana Maria Carvalho.
© 2014. John Benjamins Publishing Company
is electronic le may not be altered in any way.
e author(s) of this article is/are permitted to use this PDF le to generate printed copies to
be used by way of oprints, for their personal use only.
Permission is granted by the publishers to post this le on a closed server which is accessible
to members (students and sta) only of the author’s/s’ institute, it is not permitted to post
this PDF on the open internet.
For any other use of this material prior written permission should be obtained from the
publishers or through the Copyright Clearance Center (for USA:
Please contact or consult our website:
Tables of Contents, abstracts and guidelines are available at
John Benjamins Publishing Company
© 2014. John Benjamins Publishing Company
All rights reserved
On the partially divergent phonology
of Spanish, Portuguese and points in between
Letânia Ferreira and D. Eric Holt
Universidade Federal de Pernambuco and University of South Carolina
In this chapter we explore the comparative markedness and phonotactics of
Spanish and Portuguese phonology as a framework in which to consider: (i) the
segmental inventories of each language, (ii) the phonological processes of the
grammar of each language (including how prosody may impact these), as well as
(iii) certain historical and dialectal developments, including (iv) phonologically
intermediate/hybrid variants.
Keywords: Phonological processes, segmental inventories, syllable structure,
historical change and dialectal variation, interaction of prosody and morphology
1. Introduction1
Despite common background Portuguese and Spanish have developed several singulari-
ties since their gradual evolution from Latin. In this chapter we address these particu-
larities in phonology and seek to build parallels between these languages. Further, ob-
serving other dialects also spoken in the Iberian Peninsula allows us to construct a
richer and more complex portrait of some of the phonological characteristics of Iberia as
a whole, and to see that there oen exists a continuum of variation of dialectal features.2
ere are several factors that inuence the phonology and phonetics of a language
at the segmental level, including the position of a segment within the syllable, its so-
nority, and its placement in relation to word stress, which help us understand the lin-
guistic processes that have occurred (or not). In this chapter we adduce examples from
Portuguese (both European and Brazilian) as well as Galician, Mirandese and various
1. e following abbreviations are used in this work: Hispano-Romance (HR), Galician/
Portuguese (GP), European Portuguese (EP), Brazilian Portuguese (BP). e label GP is used to
indicate either that a given phenomenon is shared by Galician and Portuguese, or to refer to the
stage when they were unitary.
2. See also Guy (this volume) for discussion of variation and change in Latin American Span-
ish and Portuguese.
© 2014. John Benjamins Publishing Company
All rights reserved
 Letânia Ferreira and D. Eric Holt
dialects of Spanish to focus on the syllable as the domain for interaction of the linguis-
tic processes.
e chapter is organized as follows: In Section 2, we present the segmental inven-
tory of Portuguese and Spanish consonants (Section 2.1) and vowels (Section 2.2) and
provide historical background for the dierences between them, with Section 2.2 fur-
ther subdivided to treat themes including the vocalic evolution from Latin to Hispano-
Romance, characteristics of unstressed and nal vowels, as well as certain processes of
vowel harmony. In Section 3 we turn to phonological processes involving the interac-
tion of vowels and consonants, in which we address nasal and nasalized vowels and the
palatalization of dentals preceding front vowels. Section 4 is dedicated to syllable
structure and its relationship with markedness and phonotactic constraints on the dis-
tribution of both consonants and vowels. In Section 5 we address certain morphopho-
nological considerations: consonantal and vocalic alternations as well as dialectal
variation relative to number (Section 5.1), prosodic restrictions and the placement of
weak pronouns in future and conditional verbal forms (Section 5.2), and additional
issues regarding the placement of clitic forms in verbal constructions and dierences
in contractions of prepositions and articles (Section 5.3). In Section 6 we oer con-
cluding remarks.
2. Segmental inventories of Spanish and Portuguese
(with historical commentary)
2.1 Consonants
In Tables 1 and 2 below are presented the inventories of contrastive segments (pho-
nemes) of Portuguese and Spanish, beginning with consonants. Segments occurring
primarily in a regional variant are indicated between parentheses.
Table 1. Portuguese consonantal phoneme inventory (e.g., Mattoso Camara, 1970; Callou
& Leite, 1990).
bilabial labio-
dental alveolar post-
palatal velar uvular glottal
plosive p b t d k g
nasal m n ɲ
fricative f v s z ʃ ʒ
aricate (ʝ)
lateral l ʎ
rhotic ɾ, r (x) (ʀ, ʁ, χ) (h)
© 2014. John Benjamins Publishing Company
All rights reserved
On the partially divergent phonology of Spanish, Portuguese and points in between 
Table 2. Spanish consonantal phoneme inventory (e.g., Quilis, 1993).
bilabial labio-
dental alveolar post-
palatal velar uvular glottal
plosive p b t d k g
nasal m n ɲ
fricative f (θ) s ʝx (h)
aricate tʃ
lateral l (ʎ)
rhotic ɾ, r (x) (ʀ, χ)
It can clearly be seen that the two languages share large portions of their consonantal
inventory, which is understandable since both languages derive from Latin. Several
dierences must be highlighted, however, and it is not always the case that a given
sound of Spanish or Portuguese has the same sources. Note that some dialect variation
is treated in this section; other aspects (including some allophony) are treated when
discussing certain active phonological processes (Section 3) and syllable structure
(Section 4).
First, it may be noted that in certain ways, Portuguese consonantism reects a
stage closer to Latin, and one which is similar to the early common Hispano-Romance
period and from which both Spanish and Portuguese then underwent additional dif-
ferential developments.
Portuguese retains the /v/ that resulted from intervocalic voicing of /f/3 (e.g., -
 > proveito), as well as from initial /w/ (e.g.,  > [v]er); in contrast, in
Spanish, initial /w/ merges with intervocalic /β/ (from Latin /b/), and by the 1550s /β/
has merged with /b/ (from Latin /bb/ and initial /b/); Penny, 1991, p. 84–86) and the
current contextual realization of [b] or [β] is cemented.4 Likewise, Portuguese also
retains the contrast between /-s-/:/-z-/ and /ʃ/:/ʒ/ that emerges from the voicing com-
ponent of lenition.
3. is is part of the larger process of Western Romance lenition, which consisted of an interre-
lated set of changes in Late Spoken Latin consonants that includes the spirantization of the voiced
obstruents, the voicing of the voiceless series, and degemination. See Penny (1991, p. 65–67). e
process of degemination, and its relationship with the loss of a length distinction among Latin
vowels, is treated in optimality-theoretic terms in Holt (1997, 2003b) (partially discussed below).
4. is phenomenon of stop~spirant alternation that aects the voiced obstruents /b, d, g/,
shared by EP, is lacking in BP, where the stop variant predominates (Castro, 2004; Cintra, 1995;
Espírito-Santo, 2010).We are unaware of any formal treatment of these dierences, and leave the
matter for future research. ere is likewise variation in Spanish, where in certain dialects in
Central America and Colombia, /b, d, g/ only spirantize aer vowels. See Lipski (1984, p. 290),
Harris (1985), and Amastae (1995), inter alia. See also Section 4.2.1.
© 2014. John Benjamins Publishing Company
All rights reserved
 Letânia Ferreira and D. Eric Holt
Portuguese also retains initial /f/ (cf. Ptg. folha to Sp. hoja), whose probable bila-
bial pronunciation ([φ]) in Spanish territory became subject to relaxation (articulatory
undershoot) of the bilabial gesture, resulting in glottal [h], later itself articulatorily
weakened and lost ([]) in most areas by the 16th century. (For a fuller discussion,
including of the socio-historical context, see Penny, 1991, p. 79–82).
Classical Latin lacked the palatal segments (/ɲ, ʎ/), which developed in Late
Spoken Latin via palatalization of /n, l/ plus yod (palatal glide, created when hiatus was
lost, e.g.,  > [vinja] > [viɲa] viña/vinha;  > [muljer] > [muʎer] Ptg. mul-
her); /ʎ/ also resulted from palatalization and subsequent simplication of the Late
Spoken Latin (secondary) groups /k’l, g’l/ occasioned by syncope (e.g., () >
[abeʎa], () > [reʎa]). In Spanish a further transformation that occurred
(according to Martinet, 1974, p. 431; to avoid confusion with emerging /ʎ/ from – –,
see next paragraph) by the time of literary Castilian yields [ʒ] (spelled gi or j, e.g.,
mugier, abeja, reja), where /ʎ/ loses its laterality and becomes the aricate [dʒ] (e.g.,
 > [aʎo] > [a(d)ʒo] (later simplied and devoiced to [aʃo] and velarized to MSp.
[axo] ajo); more on the evolution of C+l clusters and of sibilants below).5
Spanish /ɲ, ʎ/ have an additional source, Latin intervocalic –  – and –  –, in
stark contrast with Portuguese, whose reexes of the Latin geminates are singleton /n,
l/. It has been argued (Holt, 2003b) that the simplication of the geminate sonorants is
the last step in the process of lenition from Latin to Hispano-Romance, which rst af-
fected obstruents and occurred in this order due to the gradual implementation of
restrictions on mora-bearing segments according to their inherent sonority (Zec,
1995), with more vowel-like segments retaining their mora longer. In Galician/Portu-
guese, Latin /-n-, -l-/ had been elided,6 which allows for the simplication of the gem-
inate sonorants /-nn-, -ll-/ to /-n-, -l-/; since these have disappeared from Latin, no
(or little) confusion or merger results. (See Table 3.)
5. e rst stage of this is treated in greater theoretical detail in Holt (2000), where “delateral-
ization may be seen as simplication of the doubly-articulated corono-dorsal (“palatal”) seg-
ment via loss of the feature [liquid] (not “[lateral]”); the result being a multiply-articulated
voiced palatal obstruent that is both [–cont] (at the primary coronal node) and [+cont] (at the
vocalic dorsal node), i.e., aricate [dʒ]. is approach may likewise be followed to explain those
modern Spanish dialects that have eliminated [ʎ] for aricate [dʒ] or fricative [ʒ] (> [ʃ]).
6. is too appears to be part of the process of lenition (here, the component of spirantization
involving the autosegmental spreading of [+cont] from a preceding vowel) that only aected
voiced stops /b, d, g/ in Spanish, which were sometimes subsequently lost (e.g.  > [reɣina]
> OSp. reína). at is, /n/ and /l/ would come to be [+cont], but this is a marked conguration,
and these segments are subsequently lost. See Holt (2000) (Appendix) and Holt (2002) for fur-
ther discussion of this and related issues, and Morales-Font & Holt (1997), where this loss of
/-n-, -l-/ is understood as a case of nucleation, along the lines of Colman (1983). See Section 2.1
below for discussion of the impact of the loss of /-n-, -l-/ on vowel realization, and Section 4.2
for discussion of the singular/plural alternations that arose in GP and related varieties.
© 2014. John Benjamins Publishing Company
All rights reserved
On the partially divergent phonology of Spanish, Portuguese and points in between 
Table 3. Latin intervocalic nasals and laterals and their reexes.
Spanish Galician/Portuguese Gloss
i. (/n/ < Lat. /n/) ( < Lat. /n/)
luna lua ‘m o on’
tener ter ‘to have
ii. (/l/ < Lat. /l/) ( < Lat. /l/)
cielo céu ‘sky, heaven’
palo pau ‘stick
iii. ([ɲ] < Lat. /nn/) (/n/ < Lat. /nn/)
caña cana ‘can e
año ano ‘year’
iv. ([ʎ] < Lat. /ll/) ([l] < Lat. /ll/)
bello belo ‘pretty’
castillo castelo ‘castle’
As seen in (i, ii) in Table 3 above, Latin /-n-, -l-/ were retained in Old Spanish, and this
leads to the palatalization of /nn, ll/ around the 10th or 11th century, where simplica-
tion without change would lead to merger with Latin /-n-, -l-/. As Penny (1991, p. 71–72)
suggests, simplication takes place in spite of the retention of n and l, with the resulting
phonemes “no doubt” coming to dier to preserve the distinction between /n, l/ and
simplied /nn, ll/. Lloyd (1987, p. 243) likewise suggests that speakers would be inclined
to seek to avoid phonemic merger, in this case changing articulation, which would main-
tain contrast. e additional articulatory force involved in producing geminates comes
to be realized another way, with the tongue spreading out in its contact with the alveo-
palatal region; with this palatal quality now distinguishing the simplied segments from
original /n, l/, the redundant feature (here, duration, one of the phonetic correlates of
phonological moraic weight) would be lost, leaving a system of only short segments.7,8
Another source of Spanish /ʎ/ comes from the development of Classical Latin
(primary) muta cum liquida clusters (more specically, voiceless C + l, e.g.,  > llave
[ʎ]ave,  > llover,  > llama), which yield /tʃ/ medially in Spanish
(signicantly, most frequently aer a nasal, e.g., ancho, hinchar, mancha), as well as ini-
tially in GP (chave, chover, chama) later simplied to /ʃ/ around 1700 (Williams, 1962,
7. A detailed optimality-theoretic account of these facts is given in Holt (2003b). e notion
of merger avoidance implemented follows the Dispersion-eoretic approach of Padgett (2003)
that considers not just segments themselves, but systems of segments. Note that the articulatory
doubleness of the Latin geminates (that is, originally two units of time) can be understood to be
maintained under the assumption that palatal segments are corono-dorsal structures (that is,
involving two articulatory gestures).
8. See Section for discussion of the dierential treatment of /-l/ in Spanish and
© 2014. John Benjamins Publishing Company
All rights reserved
 Letânia Ferreira and D. Eric Holt
p. 63). As mentioned above, secondary /k’l, g’l/ clusters yield /ʎ/ (e.g., () > HR
[oʎo] Ptg. olho, presumably through an intermediate unattested stage */kʎ, gʎ/ here now
further explained, and which is supported by attestations in the written record for the
primary C+l clusters.9 e emergence of /ʎ/ is understood to be due to the partial
assimilation of alveolar /l/ to the velar region of the stop that precedes it, with the ap-
proximation yielding a doubly-articulated corono-dorsal (“palatal”) segment, and this
palatalized [ʎ] then generalizing via analogy to initial -, -. In bare initial position,
these clusters are later reduced (leaving Spanish ll-), but in medial position (where they
are protected due to the assimilation between the preceding nasal and the initial stop)
they subsequently undergo devoicing and reanalysis to /tʃ/, both already present in the
rst Castilian documents. For Galician/Portuguese, it appears that the initial clusters un-
derwent a similar host of changes (palatalization, devoicing, possibly also delateralization)
to reach ch only later, though still before the appearance of the rst written records.10,11
us, from a variety of sources and due to a number of historical phonological
processes (at times, dierentially), did Spanish and GP develop the postalveolar (/tʃ, ʃ,
(d)ʒ/) and palatal (/ɲ, ʎ/) segments that were lacking in Classical Latin.
Finally, Spanish /ʎ/ was subject to further change to /ʝ/,12 probably in the period of
late Old Spanish (see Penny, 1991, p. 93); this yeísmo (the loss of distinction between
/ʎ/ and /ʝ/ in favor of articulatorily simpler /ʝ/) occurs in and is the accepted norm in
most of the Spanish-speaking world, and as Lipski (1989a, p. 211) notes, it also aects
most popular varieties of BP (Caipira and others), as well as many European dialects
and all Portuguese-based creoles (Amaral, 1955, p. 48; Azevedo, 1981, p. 79; Bortoni-
Ricardo, 1985), e.g., nonstandard abêia for abelha, and even with merger with preced-
ing /i/, as in a for lha.
An additional consonantal dierence to be discussed here involves further devel-
opments of the sibilants that lead to the emergence in Spanish of /θ/ and /x/, both ab-
sent from Portuguese.13 While GP and Spanish share the stage of dearication (before
1500) of /ts, dz/ (< Latin /t/ or /k/ plus front vowel), the results diverge. In Castilian
Spanish the resulting dental /s̺, z̺/ remain distinct from alveolar /s, z/, while Andalusian
9. Lloyd (1987) cites Upper Aragonese forms cllau [kʎ] ‘key’, pllover [pʎ] ‘to rain’, and lama
[fʎ] ‘ame.
10. Wireback (1996) discusses factors involved in the delayed spread of these changes in GP.
For overall fuller discussion, see Holt (1998, 2000, 2007).
11. Spanish /tʃ/ likewise results from the palatalization of /jt/ sequences that arise from various
sources ( > [mujto] > [mutʃo],  > HR [nojte]> [notʃe]).
12. We abstract away from the River Plate Spanish pronunciation of ll and y as one or the other
of /ʒ/ or /ʃ/ (so called ʒeísmo or ʃeísmo, with or without distinction); these are a contemporary
13. A complete discussion of the interrelated changes that aect the entire sibilant system is
beyond the scope of this chapter; the interested reader is referred to Penny (1991, pp. 86–90).
For a broad optimality-theoretic approach couched in Dispersion eory, see Baker (2003), and
for a treatment that focusses on devoicing, Bradley & Delforge (2006).
© 2014. John Benjamins Publishing Company
All rights reserved
On the partially divergent phonology of Spanish, Portuguese and points in between 
Spanish and GP appear to have lacked the alveolars, so dearication led to merger.
Subsequent devoicing leads to the situation in Castilian of having a series of voiceless
sibilants (/s̺, s, ʃ/) in precariously close acoustic space, and confusions are attested both
in commentary and misspellings of the period. is situation apparently brings about
a redistribution (16th-17th centuries) to maximize acoustic discriminability, with den-
tal /s̺/ moving farther forward away from /s/, yielding /θ/ (caza vs. casa, so called “dis-
tinción”), and /ʃ/ moving further back from /s/, yielding /x/ (bajo) in the north. In
Western Andalusia and in rural American Spanish, despite having only two sibilant
segments (vs. the more crowded three in Castilian), retraction still occurred, but yield-
ed /h/, which merged with the lingering aspirate reex of Latin /f/ (Fradejas Rueda,
2000, p. 159), that may have served as a model or attractor/target for the retracting /ʃ/.
One might wonder why Spanish /ʃ/ retracted to /x/ (or /h/), while GP did not.
Again, systemic factors appear to play a decisive role. Recall that the sources of /ʃ/ in
each language are not identical, and that this sound enters into dierential relations
with others in the respective segmental inventories. Since devoicing did not occur in
GP, /ʃ/ (deriving from Latin /sj/, e.g., HR, MPtg. baixo < )14 remained in con-
trast with /ʒ/ (< Late Spoken Latin /zj, dj, gj/, e.g., beijo, hoje, fujo < , ,
) and apparently there was no need to retract it to enhance or reinforce contrast
with /s/.15 e /ʒ/ in question here for Spanish has a dierent source, the /ʎ/ that arose
from Latin /lj, kl, g’l/, which as discussed above, became /ʒ/ presumably to avoid con-
fusion with the /ʎ/ resulting from the simplication-cum-palatalization of –  –,
which was unnecessary in GP since intervocalic /l/ had been elided. When Spanish /ʒ/
did undergo devoicing (as did /z/ > /s/), the problematic close acoustic space then re-
sulted that led to retraction of /ʃ/ to /x/ (along with the fronting of /s̺/ to /θ/).
Finally, we briey discuss rhotics, which show much variation in both Spanish and
Portuguese. While all dialects in both languages usually maintain a distinction be-
tween two rhotics intervocalically (most basically, /ɾ/-/r/, with the tap disallowed
word-initially and syllable-initially aer a consonant, and the tap-trill distinction neu-
tralized syllable-nally), there are many realizations possible. While tap /ɾ/ is usually
stable, trill /r/ oen undergoes modication in place and/or manner of articulation.
For example, there are velar and uvularized realizations in Portuguese such as fricative
/x, χ, ʁ/ and the voiced trill /ʀ/, as well as glottal fricative /h/.16 In Portugal, /r/ is
14. e modern /ʃ/ that derives from initial , , , e.g., chover, chama, chave by the simpli-
cation of /tʃ/ probably occurred later (around 1700) and so may not have had any interaction
with the opposition /ʒ/-/ʃ/ discussed here.
15. Likewise, /s/ remained in opposition to /z/, as well as /f/:/v/. e retention of /f/, rather than
the /h/ of (Andalusian) Spanish would also mean that /h/ could not serve as a model or attrac-
tor/target for GP /ʃ/ that might likewise have encouraged its retraction.
16. See Cristófaro Silva (2007, p. 143) for the distribution of the possible rhotic sounds in BP,
including several not discussed here, as well as Mateus & d’Andrade (2000, pp. 137–141), where
the syllable-nal realizations in EP and BP are analyzed in terms of delinking of autosegmental
features as well as the supplying of default values. See also Section
© 2014. John Benjamins Publishing Company
All rights reserved
 Letânia Ferreira and D. Eric Holt
typical, and /ʁ, ʀ/ are known as well. In parts of Brazil, /r/ remains, though voiceless /x,
χ, h/ have become predominant and even gained social prestige. Velar and uvular real-
izations are likewise known in several Caribbean dialects of Spanish, perhaps most
well-known being the case of rural Puerto Rican, where Puerto Rico may be realized as
[pwelto xiko].17
2.2 Vow e l s
2.2.1 Evolution of the inventory of stressed vowels
Similarly to the consonantal system, GP exhibits a set of stressed vowels (see Table 4
below) that it once shared with Spanish during the initial Hispano-Romance period,
aer the loss of contrastive vowel length in Late Spoken Latin and the emergence of the
mid lax vowels. (Unstressed and nasal vowels are treated separately below along with
other vocalic alternations.)
e principal dierence between the inventories of stressed vowels is the absence
of the open mid (lax) vowels in Spanish. When speakers reformulated the Latin Stress
Rule to favor stressed syllables as heavy, the intensied stress accent in Spanish terri-
tory due to heavier Germanic inuence18 meant that the increased duration led to
Table 4. Late Spoken Latin, Hispano-Romance, Galician/Portuguese and Spanish
tonic vowels.
(i) Late Spoken Latin, Hispano-Romance,
(ii) Spanish
i u i u
e o e o
εɔ a
a (ˈɛ > je, ˈɔ > we; ɛ̆ > e, ɔ̆ > o)
17. See Bradley (forthcoming) for a review of rhotic realizations in Spanish. e neutralization
of /l/:/ɾ/ shown in this example is (stereo)typical of Andalusian and Puerto Rican Spanish; free
variation of /l/ and /ɾ/ in initial position are known in literary Portuguese, e.g., uir (‘ow’)~fruir
(‘take pleasure in’), frecha or echa. See Mattoso Camara (1972, p. 43).
18. As discussed in greater detail in Holt (2003b), it appears that the strong accent of intensity
characteristic of Germanic (Meillet, 1970, p. 38) was slower to take hold in the more isolated
territory where GP was to develop, and several pieces of evidence support that stress was weak-
er there (Williams, 1962, pp. 11–13, 53, 56–57, 78, 87–88): there is less syncope (e.g., -ável vs. Sp.
-able); yod/wau were slower to form, as indicated by the voicing of the intervocalic consonants
in, e.g.,  > saiba vs. Sp. sepa ‘s/he know (subj.)’, and  > soube, vs. Sp. supe ‘I knew,
found out’; there is longer retention of hiatus (e.g., the syllabic value of e in forms like fêmea
‘female, from versication); and diphthongization itself (c[ɛ]u vs. cielo, f[ɔ]go vs. fuego). See also
Lleó (2003), Duell (1999) and Bayo Julve (1998).
© 2014. John Benjamins Publishing Company
All rights reserved
On the partially divergent phonology of Spanish, Portuguese and points in between 
lengthened lax vowels. ese became articulated heterogeneously ([eε, oɔ]), with sub-
sequent dissimilation leading to [je] and to [wo] (as in Italian) rst, then [we], eventu-
ally leading to awareness of the fractured articulation and the lexicalization of the
so-called alternating diphthongs /je < ‘ɛ/ (e.g., bien vs. benéco, pienso vs. pensar) and
/we < ɔ/ (e.g., bueno vs. bondad, cuento vs. contar). (In unstressed position, the lax
vowels merged with their tense/close mid counterparts, as shown in the preceding
nondiphthongal forms of the alternating pairs.)
2.2.2 Phonological processes that aect vowel realizations
ere are a number of additional complexities that characterize the vocalic systems
of GP and of Spanish. For instance, EP shows an additional stressed vowel [ɐ] that
occurs only before a palatal (e.g., telha ‘tile, lei ‘law’) or nasal consonant (e.g., canto ‘I
sing’, cama ‘bed’). Other realizations depend on nonsegmental factors. (See Mateus &
d’Andrade, 2000, Section 2.2.2 for extensive discussion of vowels.) Vowels in unstressed and nal position
ere are limitations in both languages regarding those vowels that may appear out-
side of the main stressed syllable. In Spanish, for instance, word-nal unstressed /i, u/
are uncommon (casi, tribu), except in the frequent unstressed monosyllables mi, si, ti,
su, tu, and raising of nal /-e, -o/ is common in Asturian and Galician and in the
Spanish of speakers from these regions, as well as in indigenous Andean Spanish
(probably due to Quichua inuence, where /e, o/ are absent).
For Portuguese, vowel quality and duration likewise tend to interact with stress
and word position on a very dramatic scale. According to Mateus & d’Andrade (2000)
unstressed vowels in EP are normally neutralized and are frequently deleted, where
pre-tonic and nal positions are ideal for vowel reduction and deletion, especially in
colloquial speech, e.g., precisão ‘precisionas [pɾ.si.ˈzãw̃], bato I beat’ as [ˈbat] or [ˈba.
tu], and dever [ˈdver] (vs. devo [ˈ]). e tense mid vowels /e, o/ tend to be raised
when in unstressed position (e.g., meninoboy’ [mĩ.ˈnĩ.nu], gordura ‘fat’ [gur.ˈdu.ɾa]),
and in BP, especially in nal position (e.g. [ˈ] sabe ‘s/he knows’, [ˈʒu.ɾu]), juro I
swear’ and as a result the most common word-nal vocalic realizations in BP are /i, a,
u/ (Callou & Leite, 1990).19 (Other reduced vowels include [ɨ, ə, ɐ].) Words bearing nal
stress, however, maintain the lax realization, e.g., [ka.ˈfɛ] café ‘coee’, [a.ˈvɔ] avó ‘grand-
mother’. Callou and Leite (1990) argue that this tendency towards vowel raising is not
stigmatized and is found in a number of sociolects of BP, adding that speech rate is a
factor as well, while Vieira (2002) shows that for this dialectal variation, the type of syl-
lable and segment preceding the target vowel can inuence vowel height realization.
e group of mid vowels is indeed quite a rich environment for phonological
processes in the languages originally from Iberia. Parkinson (1988) points out that
19. Additionally, the reduction to [i] feeds the process of palatalization of a preceding coronal
consonant, e.g., [si.ˈda.dʒi] cidadecity’; see Section 3.2 for discussion.
© 2014. John Benjamins Publishing Company
All rights reserved
Letânia Ferreira and D. Eric Holt
the contrast between GP high-mid and low-mid vowels (/e/:/ɛ/; /o/:/ɔ/) is frequently
neutralized, as they historically were in Old Spanish. Alves (1999) has found that in
Minas Gerais, Brazil, some words can be pronounced with either of the mid vowels,
e.g., corpete ‘corset’ as corp[e]te or corp[ɛ]te, where, the author suggests, frequency of
usage can inuence the choice made by the speaker. Several morphophonemic alter-
nations, discussed in the following section, are also related to this vocalic interchange
([e~ɛ]; [o~ɔ]). Vowel harmony processes.
Vowel harmony involves a restriction on the vowels that may co-occur in a given do-
main, with typically the nal vowel triggering the sharing of some feature (height,
roundness, tenseness/laxness, etc.) among preceding vowels (Penny, 2009). Metaphony
is the term usually applied to Romance versions of this process, and it occurs both in
Iberia and the Americas, though not always to the same degree or in the same way, and
with varying social prestige.
Historically, Old Spanish underwent a process of raising of mid vowels by inec-
tion of a yod in the nal syllable, where the presence of the palatal glide causes /e, o/
to take on its [+high] feature.20 is phonological process applied regardless of
grammatical category (that is, not just to verbs, though these are the most oen cited
forms), e.g., dormir~durmió ‘to sleep’ (innitive vs. 3rd sg. pret.), sentir~sintió ‘to
feel’ (innitive vs. 3rd sg. pret.), llover~lluvia ‘to rain vs. rain’, and is now merely a
historical lexical relic. Additionally, historically Old Spanish showed an alternative
(and at the time, even more frequent) set of conditional endings, –ié(n), whose yod
likewise triggered assimilatory raising of the preceding vowel (cf. MSp. decían vs.
OSp. dizién).
e metaphony of Portuguese, however, is an active and more widespread pro-
cess21 and is generally morphophonological in nature. Nunes (1989) says that the dif-
ference in openness between the mid-low and the mid-high vowel in Portuguese is
related to number agreement (e.g., ovo/ovos [ˈovu/ˈɔvus] ‘egg/eggs’), and gender-related
alternations also show changes in vowel quality (e.g., novo/nova [ˈˈnɔ.va] ‘new
(m.)/(f.)’). According to Cagliari (1997) these cases are marked in the lexicon of the
language, and they are the result of “a rule of vocalic dierentiation” in which the mas-
culine forms are composed of two mid-high vowels and the feminine forms are com-
posed of two mid-low vowels. Dierent from Asturian and Portuguese metaphony,
however, Galician metaphony does not work as a number-contrast marker (Penny,
20. Yod in fact has various sources as well as exerts various phonological inuences historically,
a full discussion of which is well beyond the scope of this chapter. See Penny (1991), inter alia.
21. Penny (2009) has established a parallel between Italian metaphony and that found that in
Iberian dialects, and Cantabrian, Asturian and Galician do present several cases of metaphony
that are quite similar to that found in Portuguese.
© 2014. John Benjamins Publishing Company
All rights reserved
On the partially divergent phonology of Spanish, Portuguese and points in between 
2009), where the raising of vowels occurs in both singular and plural forms: [’so.gro/’so.
gros] ‘father(s)-in-law’ vs. [’sɔ.gra/’sɔ.gras] ‘mother(s)-in-law’.22
3. Phonological processes involving vowels and consonants
3.1 Nasal(ized) vowels
When a vowel occurs next to a nasal consonant (/m, n, ɲ/), in both GP and Spanish it
undergoes nasalization, an assimilation process that relates to the (mis)timing of the
opening of the velic port required for the articulation of surrounding nasal conso-
nants, e.g., canta ‘(s/he) sings’ /kan.ta/ as [ˈkã.ta] (GP), [ˈkãn.ta] (Spanish). Despite
undergoing the same process, GP nasalized vowels present a much stronger degree of
nasality than Spanish ones (Masip, 1999, p. 53),23 as might be inferred from the lack of
nasal consonant in the GP surface form.
e more striking dierence between GP and Spanish is the presence in the for-
mer of the so-called nasal vowels ([ĩ, , ɐ̃, õ, ũ]), with indication in the orthography
with tilde only for ã, õ and nasal diphthongs ([ɐ̃j̃, õj̃, ũj̃, ɐ̃w̃]), which reects a historical
intervocalic /-n-/ that has been lost (discussed in Section 2.1), and where the nasaliza-
tion may reect a surface contrast (e.g., pão [ˈpãw̃] ‘bread’ vs. pau [ˈpaw] ‘stick, ‘wool’
vs. ‘there’). e status of these nasal vowels (either as independent phonemes or as
the surface realization of an underlying sequence of vowel + nasal consonant, perhaps
underspecied/archiphonemic /N/) has been the subject of much debate,24 and while
speakers may be unaware of the consonantal feature in nasal vowels (Sampson, 1999),
acoustic studies show that nasal vowels are indeed composed of two dierent parts, a
vowel plus a nasal coda. For the phonological alternations between singular and plural
nasal-nal forms, see Section 5.1.2.
3.2 Palatalization of /t, d/ + front vowel in BP
Present in several dialects of BP is the palatalization of dental consonants triggered by
front vowels that results in aricates [dʒ] and [tʃ], e.g., [ˈʒen.tʃi] gente ‘people, [si.ˈda.
22. For formal approaches to vowel harmony in Portuguese, see Bisol (1989), Walker (2005,
2011), Wetzels (1991, 1995), and many others. Vowel harmony processes are also common dia-
lectally in Spanish and related varieties, and have also been treated formally (in phonological and/
or morphological terms), like that of Eastern Andalusian Spanish (Zubizarreta, 1979; Walker,
2005; Jiménez & Lloret, 2007), Pasiego/Montañés (Penny, 1969; McCarthy, 1984), and Lena Bable
(Hualde, 1998; Holt, 1999; Martínez-Gil, 2007), as well as Puerto Rican raising (Oliver, 2008).
23. See Callou & Leite (1990) and Lacerda & Rossi (1958). Further, Frota & Vigário (2000)
show that between BP and EP, nasal vowels tend to be much longer than oral ones in BP when
compared to the corresponding vowels in EP.
24. See Lipski (1985), Medeiros (2011), Mateus & d’Andrade (2000), Morais-Barbosa (1962), Mo-
rales-Front & Holt (1997), Parkinson (1983), Reed & Leite (1956), and Shosted (2006), inter alia.
© 2014. John Benjamins Publishing Company
All rights reserved
Letânia Ferreira and D. Eric Holt
dʒi] cidadecity’. In dialects such as the one spoken in Recife this palatalization is vari-
able, such that dierent individuals might not apply the rule/process, and an individ-
ual speaker might apply it variably, with tiaaunt’ realized as either [ˈti.a] or [ˈtʃi.a]. is
process is independent of word stress, and we nd examples where both tonic and
atonic front vowels trigger the process: [ˈdʒi.ka] dica tip’, [dʒi.ˈɲej.ɾu] dinheiromoney’,
[ˈɔ.tʃ] ótimo ‘great, optimal’.
4. Syllable structure, markedness and phonotactics
In addition to some dialectal variation in the realization of contrastive segments
(especially among liquids), there are also contextual processes that lead to additional
segmental allophony. In this section, we explore syllable structure and the role it plays
in the distribution of segments and of certain segmental realizations due to their posi-
tion in the syllable and to the restrictions imposed on these structures. (Vocalic real-
izations, conditioned by dialectal and prosodic factors, as well as vowel harmony, were
treated in Section 2.2.2.)
4.1 Syllabic structure
As highlighted by Colina (2009), syllable structure has played an important role in
phonological theory, because it is in this domain that many of the phonological phe-
nomena that aect segments will manifest themselves. Syllables tend to follow a cer-
tain organizational hierarchy based on sonority when grouping segments (see Selkirk,
1984; Clements, 1990; inter alia), and are composed of an optional onset and an oblig-
atory rime, with obligatory nucleus and optional coda. In turn, the phonological gram-
mar of a language will impose restrictions on each of these structural positions. Each
of the syllabic elements (onset, nucleus, coda) may be maximally branching, therefore
allowing for the possibility of onset clusters, diphthongs, and to a lesser extent, com-
plex codas. When complex, the onset tends to abide by minimum sonority dierences,
and the coda faces greater limitations and will more commonly undergo allophonic
processes as well as weakening and deletion of segments that normally work towards
syllable openness and overall lesser markedness.
While Portuguese and Spanish share many of these phonotactic constraints, allow
many syllable types and both prefer CV syllables, there are noteworthy dierences.
4.2 Onsets and onset clusters
Any of the contrastive segments of Portuguese or Spanish may occupy syllable- and
word-initial position (except /ɾ/ word-initially; and /ɲ/ is rare here). When the onset is
complex, however, there are strict limitations on the consonants that may cluster, and
© 2014. John Benjamins Publishing Company
All rights reserved
On the partially divergent phonology of Spanish, Portuguese and points in between 
these have to do with following sonority sequencing conditions and respecting a min-
imum sonority dierence between the two segments. e result is that in both Spanish
and Portuguese, plosives /p, t, k, b, d, g/ + liquids /l, ɾ/ (but not /r/ or other rhotic, nor
in most dialects the sequences */tl, dl/, excluded for reasons of complexity) may form
a complex onset, as may the labio-dental fricatives of each language form the rst
member of a cluster as well (, fr in both Spanish and Portuguese; also vr in Portu-
guese, e.g., palavra). All other consonants that enter into contact are separated across
a syllable boundary (e.g,,,
Spanish and BP agree fully in this regard; EP, however, shows additional onset
clusters that violate these sonority conditions. EP allows word-initial clusters of
plosive+plosive (e.g., ptério), plosive+fricative (e.g., psicologia), plosive+nasal (e.g.,
gnomo, pneu), as well as these and many more that result from the deletion of un-
stressed vowels (see also Veloso, 2002): espaço [ʃpasu], estar [ʃtaɾ], eslavo [ʒlavu],
pequeno [pkenu], depender [dpendeɾ], terreno [tʀenu], even resulting in phonetic clus-
ters of up to six consonants, e.g., telefone [tlfɔn], despregar [dʃpɾgaɾ], desprestigiar
[dʃpɾʃtiʒjaɾ]. Given the massive violations of otherwise robust generalizations on syl-
lable structure, Mateus & d’Andrade (2000, Section postulate empty syllabic
nuclei for EP, both for those cases that result from vowel reduction/deletion as well as
for the word-initial cases.25
For BP, when these clusters arise word initially, there is a tendency of epenthesis,
such that psicologia will be realized as [ʒi.a], and pneu will be realized as
[pi.’new],26 and this tendency to produce a vowel between consonants is also typical of
EP children during the process of language acquisition. Likewise, the syllable-nal
voicing assimilation that we observe in forms like eslavo [ʒlavu] fails to obtain for these
phonetic clusters (e.g., pneu, obter). Mateus & dAndrade (2000) take these facts to sup-
port their claim of empty nuclei for EP, such that the initial member of these phonetic
clusters is phonologically in onset position, where licensed.27
4.2.1 Phonological processes aecting onset position
e onset is by nature a strong and stable position, especially when compared to the
coda, whose markedness makes it subject to various weakening processes. One case of
onset allophony is the strengthening in Spanish of syllable-initial glides. e palatal
25. A similar case in Spanish is that of the colloquial pronunciation of estar, as in Estoy/Está
bien as [‘toj, ‘ta ‘βjen]; it appears that the reduction of initial [e] (or perhaps the suppressing of
the insertion of it before an illicit /sC/ cluster) then leads to the deletion of /s/, presumably be-
cause Spanish does not allow empty nuclei.
26. is epenthesis in BP also occurs in syllable- and word-nal position (e.g., picnic > pikiniki)
27. Vowel reduction and deletion, especially in contact with voiceless consonants (e.g., ocina
as [ofsna] or pues as [ps]) are also known in some regional varieties of Spanish, such as that of
Central Mexico, as well as highland/Andean varieties (where doce and dos can sound homopho-
nous). See Lipski (1984).
© 2014. John Benjamins Publishing Company
All rights reserved
 Letânia Ferreira and D. Eric Holt
glides may be realized with either a fricative or stop (e.g., hielo ‘ice’ [‘ʝe.lo]~[‘ɟe.lo]) and
the labiovelar glide may also undergo fortition in syllable onset position (e.g., huevo
egg’ [we.βo]~[gwe.βo]), both especially aer a pause or a nasal consonant, contexts
that likewise condition the realization of the voiced stops ([b, d, g]~[β, ð, ɣ]) (see also
fn. 4). A case of onset weakening is that of the dearication of /tʃ/, which in Andalu-
sian and some varieties of Latin American Spanish may be pronounced as [ʃ]. (See
Penny, 1991, p. 86–90 for discussion of historical dearication, and also fn. 14.)
Some onsets in BP also undergo allophony, sometimes variably. Several voiced
fricatives in BP are sometimes produced as aspiration in casual speech (Palácio, 1989).
Post-alveolar /ʒ/ may be aspirated in word-initial intervocalic position (e.g., a gente
‘we’ [a.‘hen.ti]), the alveolar /z/ in word-medial has been recorded as well (e.g., exem-
plo ‘example’ [i.’hem.plu]), and perhaps most commonly, labio-dental /v/ may also be
aspirated both word initially and medially (e.g., vamos ‘let’s go” [’], (es)tavawas’
[’ta.ha]. As discussed in Section 3.2, dental /t, d/, which occur primarily in onset posi-
tion, undergo variable palatalization before front vowels in BP.
Finally, the phenomenon of intervocalic voicing of (word-)nal s (see Table 5) was
common to Hispano-Romance and is maintained in Portuguese, and while lost in Old
Spanish, similar processes also occur in Judeo-Spanish and in Ecuadoran Spanish (see
Bradley & Delforge, 2006 and Lipski, 1989b, inter alia).
is voicing is not restricted to morphophonemic alternations and commonly ob-
tains in connected speech as well, e.g., Estados Unidos ‘United States’ as [is.’ta.du.zu.’ni.
duʃ] and os homens ‘the men’ as [u.zo.mi],28 in these cases arguably improving percep-
tion of the fricative phoneme (Ferreira & Pérez-Leroux, forthcoming) and aiding in
the identication of the morpheme- or word boundary.
4.3 e syllable rime29
While in Latin any consonant may appear in the coda, and /s/ may follow this conso-
nant, Portuguese and Spanish tended to reduce closed syllables through palatalization,
Table 5. Intervocalic voicing in Portuguese.
Spanish Portuguese
sg. pl. sg. pl.
mes (mês) ‘month [ˈmes] [ˈ] [ˈmejʃ] [ˈme.ziʃ]
feliz ‘happy’ [fe.ˈlis] [fe.ˈ] [fe.ˈliʃ] [fe.ˈli.ziʃ]
28. In the latter example, the plural marker of the noun undergoes weakening, while that of the
article is maintained due to its resyllabication in onset position. See Ferrari Neto (2003) and
Corrêa et al. (2005) for discussion of BP’s tendency to maintain the plural marker in articles.
29. is section draws largely on Morales-Front & Holt (1997, pp. 426–431, Appendix).
© 2014. John Benjamins Publishing Company
All rights reserved
On the partially divergent phonology of Spanish, Portuguese and points in between 
Table 6. e syllable structure of Modern Portuguese and Spanish.
Rime *C
/\ /\
/\ |
(m)a j s
arication, simplication of geminates and consonant groups, vocalization, and other
processes (Holt, 1997; Lloyd, 1994; Mattoso Camara, 1972, p. 46; Penny, 1991; Porzio
Gernia, 1976, p. 210–211; etc.). Consequently, in both Modern Portuguese and Modern
Spanish, coda plosives (e.g., apto) and consonant clusters (e.g., obstrucción/obstrução)
are now limited, mostly having been restored through learnèd inuence (see espe-
cially the works cited above by Mattoso Camara & Penny), and likewise undergo re-
duction/elimination in casual speech (e.g., Sp. do(c)tor, se(p)tiembre, tra(n)sporte);
consonants closing the syllable are now primarily limited to the sonorants r, n, l and s,
with Spanish also allowing /θ/ (in Castilian Spanish) and occasionally /x/ (e.g., reloj).
is yields the syllable template shown in Table 6, which shows that the rime may
branch, but not the coda.
4.3.1 e syllable nucleus, diphthongs, hiatus and rhythm
In both Portuguese and Spanish, any vowel (at least in stressed position; see Section
2.2 for further discussion of vowel phenomena) may occupy the syllable nucleus, and
diphthongs are common. In Spanish, the predominant type of diphthong is rising
(GV) and vowel quality is unrestricted; in Portuguese, however, the predominant
diphthong is of falling sonority (VG), whose rst member may likewise be of any qual-
ity; rising diphthongs are unstable in Portuguese, and may always be pronounced in
hiatus (Mattoso Camara, 1972, p. 55; Peixoto, 2011), e.g. el as [‘ew] or [’ew].30 Ad-
ditionally, while Spanish has only oral diphthongs, Portuguese also has (surface) nasal
diphthongs, always VG: e.g., [ãw̃] pão ‘bread’, [ãj̃] mãe mother’, [õj̃] põe ‘put (3p. sg.)’.
Finally, as noted above, one major dierence between Spanish and Portuguese is
related to the behavior of the syllable nucleus in EP, which tends to delete unstressed
vowels in casual speech and display empty nuclei (e.g., dever [dver] ‘to owe’, devedor
[dvdor] ‘debtor’), in contrast to BP, which like Spanish only allows vowels in nuclear
position. Indeed, the dierent way in which the nucleus can operate in these varieties
of Portuguese will result in opposite tendencies in what concerns syllabic structure,
and concomitantly, the perception of rhythm. EP speakers tend to generate closed syl-
lables and complex consonant clusters (at least phonetically) when unstressed vowels
are deleted, while BP speakers will tend to insert vocalic segments into standard
consonant clusters, thus generating an additional syllable in words and reinforcing a
high tendency towards syllable openness (e.g., abduzir [’zir] ‘to abduct’). ese
30. Additionally, as discussed in fn. 18, historically, Portuguese retained Latin hiatuses longer.
© 2014. John Benjamins Publishing Company
All rights reserved
Letânia Ferreira and D. Eric Holt
phenomena certainly impact the perception of rhythm and are the main reasons that
separate these Portuguese varieties from the rhythmic behavior of Spanish. BP is de-
scribed as a mix between a syllable- and mora-timed language variety, while EP be-
haves more like a mix between a stress- and a syllable-timed language (Frota & Vigário,
2000). Spanish, which by default does not present the traits described above that aect
regular syllable structure, is more commonly identied as a syllable-timed language.
4.3.2 Codas, coda clusters and coda modication and simplication
Returning to consonantal issues, we see that the coda faces greater limitations, both in
terms of the inventory of contrastive segments that may occur there as well as the pro-
cesses that aect this position, all of which leads toward a more optimal sonority pro-
le and nearer approximation to ideal CV structure (Clements, 1990).
Regarding the contrastive segments that occur in syllable-nal position, as stated
above, the consonants that close the syllable are now primarily limited to the sonorants
r, n, l and s, with Spanish also allowing /θ/ (in Castilian Spanish) and plosives (/p, t, k,
b, d, g/) and other fricatives (/f, x/) occurring in relatively few words primarily of lear-
nèd inuence. In Portuguese, the nal consonants can be analyzed as underspecied
/L, R, S/, all arguably [+cont] and whose values for place of articulation (and for /S/,
also voicing) are determined contextually (and dialect specically) at the post-lexical
level (Mateus & dAndrade, 2000, Section 3.2.3 and Section 7.4).31,32
Most signicantly in terms of dierences, Spanish also licenses nasal consonants in
coda position, both word medially (e.g., canciones, contestar) and nally (e.g., son, con,
bailan). In Portuguese, in contrast, nasal coda consonants are disallowed and will con-
sistently be vocalized (e.g., word nally, cam ‘they stay/remain[.kãw̃], comem ‘they
eat’ [ko.mj̃], and word medially, canto ‘I sing’ [kã.tu]). (See Sections 2.1, 3.1 and 5.1.2
for further discussion of the nucleation of nasal consonants and their realization.) Phonological processes aecting coda position.
Many are the phonological processes that target the coda, as this position faces more
limitations than onsets, and these normally yield open syllables or structures closer to
the type CV. In the evolution from Latin to Portuguese and Spanish closed syllables
eroded through palatalization, arication, simplication of geminates and consonant
groups, vocalization, and other processes.
31. Due to the tendency to delete unstressed vocalic segments in EP in word-nal position, any
consonant can be found here in EP, though Mateus & d‘Andrade (2000) analyze these cases as
containing an empty nucleus and so not in coda position in phonological terms; this phenom-
enon does not apply to BP.
32. e archiphoneme /S/ in syllable-nal position represents the neutralization of dental /s, z/
and post-alveolar /ʃ, ʒ/, which contrast only word initially and intervocalically. Voicing assimila-
tion of /S/ occurs in both Portuguese (e.g., felizmente with [ʒ]) and Spanish (e.g., esbelto with [z])
and also takes place across word boundaries (e.g., Ptg. os dias, Sp. los días). e most common
allophones for this Spanish fricative are [s, z].
© 2014. John Benjamins Publishing Company
All rights reserved
On the partially divergent phonology of Spanish, Portuguese and points in between 
One example is the case of /-l/ syllable nally, a segment that behaves dierently
in Spanish and Portuguese. Most notably, /-l/ is velarized ([–ɫ]) in EP, and is further
vocalized to [–w] in BP (e.g., Brasil [braziw] vs. brasileiro with [l]). Evidence for this
velarized pronunciation even in Old Portuguese is the use of ll syllable- or word -
nally (malldade, mortall; Williams, 1962, p. 26). e reasons for velarization may again
be attributed to the relationship between sonority and the ability to bear a mora. at
is, as the sonority restrictions of mora-bearing segments are further limited (with ob-
struent geminates already being simplied, and nasal and lateral geminates also sim-
plifying into new palatal segments), the production of /l/ with more prominent dorsal
involvement means that the secondary vocalic place of articulation is enhanced, and
more vowel-like segments carry higher sonority, and thus may be moraic. e ultimate
vocalic realization of /-l/ as [–w] in BP has lost the original primary consonantal ar-
ticulation and reects a later stage at which no consonants bear moras.33,34
Another example is the tendency towards deletion of rhotics in word-nal posi-
tion, especially for the verbal marker of the innitive. According to Callou & Leite
(1990) in this position the absence of the rhotic segment alternates with the presence
of aspiration by glottal or laryngeal frication (e.g., estar ‘to be’ [is.ˈtah]~[is.ˈta]). is
loss of innitival –r is also known in Andalusian and lowland dialects of Latin American
Spanish, where syllable-nal liquids are also frequently neutralized (e.g., falta with [ɾ],
arma with [l]). Similarly, velarization of nal nasals in these same dialects yields [-ŋ]
(e.g., pan ‘bread’ [paŋ], sometimes even leading to total loss of the consonant with
nasalization remaining on the vowel, con ‘with’ as [kõ]).
Finally, let us consider syllable- and word-nal s. is segment undergoes aspira-
tion and deletion widely in many dialects of Spanish. is weakening of the fricative is
very frequent in colloquial styles while formal speech shows higher rates of retention
(Hualde, 2005). In Portuguese, aspiration and deletion of /S/ have also been reported
(Azevedo, 1984; Carvalho, 2006, Scherre, 2001; Naro, 1981; Scherre & Naro, 2006).35
33. is is the approach taken in Holt (2002) for the vocalization in Hispano-Romance of /-l/,
whose higher sonority (as a liquid) should have meant it was exempt from the then current ero-
sion eects that vocalized other syllable-nal segments like /-k, -g/ (e.g.,  > autor). Penny
(1991) and others (including Holt, 1997, 1999) assume that /-l/ was velarized, and stipulated that
velars vocalized (e.g.,  > muito). However, by synthesizing the insights of the articulator
group hypothesis (Padgett, 1991,1995) and of the feature geometry of liquids (Walsh-Dickey,
1997), because /-l/ is seen as [-cont] (and thus low sonority) at its primary coronal articulation
it is indeed targeted by lenition. See also fns. 5, 6, as well as Mateus & d’Andrade (2000, p. 27).
34. For further discussion of /-l/, including of its gliding to [–j] in plural forms like fáceis, ho-
teis, see Morales-Front & Holt (1997, Section 4.4.1) and its fn. 12, where restrictions on mid
vowel reduction in atonic syllables closed by glides as well as /n, ɾ, l/ come together as all involv-
ing complex nuclei. See also Brandão de Carvalho (1988).
35. To the best of our knowledge, most of the Portuguese research conducted focuses on plural
markers; for information about reports in BP of /S/ aspiration not related to number agreement,
see Ferreira (2001).
© 2014. John Benjamins Publishing Company
All rights reserved
Letânia Ferreira and D. Eric Holt
Lastly, there is a tendency to palatalize nal –s, and in BP there is oen concomitant
insertion of palatal glide [j] (e.g., mês [‘mejʃ], arrozrice’ [axojʃ], Albano, 2001; Barbosa
& Albano, 2004).36
5. Some morpho(phono)logical considerations37
5.1 Singular-plural alternations
While the basics of pluralization in Spanish and Portuguese are the same (add /-s/ to a
vowel-nal word and /-(e)s/ to a consonant-nal one), Portuguese shows additional
alternations for words ending in various consonants. Some of these phonological pro-
cesses have been treated in other sections, but are collected here.
5.1.1 Consonantal and vocalic alternations
Portuguese retains the Hispano-Romance (and Old Spanish) voicing alternation in the
plural for those words whose singular form ends in /-s/: mes [‘mejʃ], meses [‘me.ziʃ]; feliz
[fe.ˈliʃ], felices [fe.ˈli.ziʃ]. Additionally, words that (at least historically) end in /-l/ show
loss in the plural: papel [pa.ˈpeɫ] or [pa.ˈpew] vs. papeis [pa.ˈpejʃ]; nal [.ˈnaɫ] or [.ˈnaw]
vs. nais [.ˈnajʃ]. e loss of /-l-/ (and /-n-/) in GP was treated in Section 2.1 as a case
of nucleation due to lenition, and the creation of nasalized vowels was treated in Section
3.1, but further dialect variation of the plurals of words with nasals is presented below.
5.1.2 Dialect variation of nasal singular-plural alternations38,39
Loss by nucleation of /n/ in GP brings about morphological alternations between sin-
gular and plural forms, in contrast with plural formation in Spanish. Consider the re-
sults of the singular and plural reexes of Latin nouns ending in –, –, and
– shown in Table 7:
36. Palatalization of nal fricatives has been noted in Manaus, Rio de Janeiro, Alagoas,
Pernambuco, Ceará and Santa Catarina (Marins & Margotti, 2012). Despite normally being
documented for lexical forms (e.g., mês ‘month[ˈmejʃ]), mas ‘but’ [ˈmajʃ]), impressionistically,
this vocalic insertion also seems to occur in at least some instances of plural formation (e.g.,
cajá/cajás ‘type of tropical fruit’ [ka.ˈʒa]~[ka.ˈʒajʃ]), but further research is needed to determine
the precise nature and extent of the phenomenon.
37. See O’Neill (this volume) for a discussion of verbal morphology not only in Portuguese and
Spanish, but also Aragonese, Mirandese and Galician.
38. Portuguese pluralization, especially of consonant-nal words, shows many complexities,
and a complete treatment is beyond the scope of this work. See Morales-Front & Holt (1997) for
an optimality-theoretic approach.
39. is section draws largely from Holt (2000), where the data are analyzed in optimality-
theoretic terms. See also Colina (2011). For additional data and discussion, see also Moraes
Ferreira (1898), Leite de Vasconcelos (1900), and Vásquez Cuesta & Mendes da Luz (1980).
© 2014. John Benjamins Publishing Company
All rights reserved
On the partially divergent phonology of Spanish, Portuguese and points in between 
Table 7. Plural formation in Spanish and Portuguese.
Portuguese Spanish Latin
irmão(s), irmã(s) hermano/a(s) (< Lat. / ‘brother/sister’)
pão ~ pães pan ~ panes (< Lat.  ‘bread’)
razão ~ razões razón ~ razones (< Lat.  ‘reason’)
e situation in other varieties closely related to Spanish and Portuguese, however,
shows additional complexities. For example, Galician shows much regional variation
(see the sets of boxed forms in Table 8; note that -n = [–ŋ]), and Mirandese (a Leonese
variant spoken in and around Miranda do Douro in Portugal) exhibits forms (Table 9)
between Spanish (with intervocalic /-n-/ retained and the plural always ending in -es)
and Portuguese (where singular form shows loss of /n/ only syllable nally with nasal-
ization of preceding vowel)40:
Table 8. Galician dialectal data (Perez, 1982, p. 209) (Reexes of Latin ,
 ‘brother, sister’).
masc. fem.
sg. irmão irmã irmán irmán
pl.irmaõs irmãs irmáns irmáns
irmao irmá irmán irmán
irmaos irmás irmás irmás
irmao irmán irmá ir
irmaos irmáns irmás irmás
irmao irmán irmán irmá
irmaos irmás irmáns irmás
Table 9. Mirandese singular-plural alternations.
sg. pl. cf. Ptg.
panes pão, pães
biẽbienes bem [bẽj̃], bens [bẽj̃ (n)s]
melõ melones melão, melões
nes m, ns: [fĩ, fĩ(n)s]
èl tẽeilles tênan (= [–ã#]) ele tem, eles têm
cf. mano mão
cheno cheio
40. See also Cristina Martins (this volume) for the “frontier status” of Mirandese.
© 2014. John Benjamins Publishing Company
All rights reserved
Letânia Ferreira and D. Eric Holt
e dierences exhibited in these two charts are due to regional variation in the im-
portance of phonological constraints on lenition (spirantization), nasal nucleation,
vowel nasalization, coda conditions and other markedness conditions, as well as to
dierential reanalysis of phonetic forms as underlying ones.
5.2 Dierences in morphological and prosodic status of future
and conditional verbal endings
While morphosyntactic phenomena are not ‘strictly speaking’ the focus of this chap-
ter, there are points of intersection with prosody. One example is the desinences of
future and conditional verbal forms of Old and Modern Spanish and GP, which de-
veloped from the present tense of the verb  ‘to haveand the imperfect tense
of either  or  ‘to go’. Over time these lose their morphological indepen-
dence and become reanalyzed as forming a unit with the innitive, with concomitant
reanalysis into a single prosodic domain for stress assignment and the innitival (now
verbal base) no longer bearing stress. In earlier stages, this process of grammaticaliza-
tion (along with the conversion of the Latin demonstratives into denite articles,
forms which did not exist in Latin) is still underway and the morphological status and
prosodic independence of these “endings” is ambiguous. ere are many attestations
of future and conditional forms still separated from their innitival base with the
atonic pronouns in between (so called medioclisis or mesóclise): e.g., amar lo é ~ lo
amaré (Lloyd, 1987, p. 311), ferlo ia (Penny, 1991, pp. 205–206), and excusarse ía
(Gracián, Criticón, from the Golden Age, cited in Lapesa, 1986, p. 392). Such forms
still obtain in EP and in formal styles of BP (e.g., cf. Sp. lo dará, lo daría vs. Ptg. da-
lo-á, da-lo-ia).41
5.3 Dierences in behavior of weak pronouns and denite articles
Two related issues are the status and placement of the new atonic object pronouns,42
and the contraction with certain prepositions of the new denite articles.
41. e interaction of the grammaticalization of the atonic pronouns, the new verbal forms and
metathesis in Old Spanish is treated in Holt (2004). Beginning in the 17th century, coinciding
with the founding of the Real Academia Española de la Lengua (1713/1714) (whose focus is to
maintain the purity of the language), metathesis no longer obtains.
For formal morphological accounts of mesoclisis in Old Spanish future constructions, see
Bouzouita (2011); for EP mesoclisis, see Luís & Spencer (2005); and for an account of nonstan-
dard Spanish plural verbal forms like de-me-lo-n for dénmelo, see Manzini & Savoia (2011).
42. See Ana Maria Martins (this volume and references cited therein) for additional exempli-
cation and discussion of the categorial status (head vs. phrase, and possible change over time)
and placement of clitics in (both Old and Modern) Portuguese and Spanish (related to the
Tobler-Mussaa law, the constraint against clitics appearing in initial position).
© 2014. John Benjamins Publishing Company
All rights reserved
On the partially divergent phonology of Spanish, Portuguese and points in between 
5.3.1 Dierences in placement of weak pronouns
Clitics are weak elements generally accepted to be structurally ‘decient’ at the sub-
lexical level of syntactic structure; as a result, they cannot introduce new referents into
the discourse, cannot be coordinated, and cannot be modied (Ana Maria Martins,
this volume). Lacking word stress, they are prosodically dependent words. Other char-
acteristics include their neutrality in stress placement in both Portuguese and Spanish
(Sp. compra/cómprame/cómpramelo, Ptg. falávamos/falávamos-lhe), and, in Portuguese,
their participation in the vowel raising processes that occur in unstressed nal sylla-
bles of lexical words (se > s[i], pro > pr[u]). In what concerns the clitic group the litera-
ture also shows prosodic dierences between BP and EP related to the intonational
contour. When evaluating the alignment of phrasal tones relative to the le edge of the
intonational phrase, Frota (2003) found that in EP this tone would always occur with-
in the domain of the rst prosodic word regardless of the number of syllables in the
word. Ferreira (2008, 2011) shows that in BP when the rst prosodic word is short, the
phrasal tone will tend to occur within the clitic group preceding this word. e access
of the phrasal tone to the clitic group in BP and not in EP suggests a stronger prosodic
boundary between the clitic group and the prosodic word in EP. (See Armstrong &
Cruz, this volume, for additional discussion of the intonational phonology of Peninsu-
lar Spanish and European Portuguese.)
With respect to placement of clitic pronouns, Spanish, EP and BP show divergent
behavior: in Spanish, the weak object pronouns occur before the conjugated verb
(lo hace) or attach to the imperative (hazlo); with innitival complements (esperaba
saludarte) or other verbal constructions, these may precede the auxiliary or be at-
tached (e.g., innitive: va a hacerlo; gerundive: está haciéndolo).43 In EP, the clitics
appear before or aer the verb (with proclisis triggered by negation, quantiers, wh-
phrases, contrastive focus and certain adverbs), and in verbal periphrasis, follow the
innitive or gerund, while in BP, they usually appear before the main verb: Eu esperava
ver-te (EP) vs. Eu esperava te ver (BP), Ela manda-lhe um livro vs. Ela lhe manda um
livro, Ele vai levar-nos vs. Ele vai nos levar, and even initially (Me empresta esse livro,
por favor), ungrammatical in EP.44
A full review of formal approaches to these and related phenomena is beyond the
scope of this chapter, but some accounts are reviewed here. Simioni (2008) argues that
prosodization of clitics is determined in an optimality-theoretic approach by the
interaction of constraints on morphosyntactic alignment and prosodic dominance
43. In the historical emergence of these forms, there was a strong tendency towards enclisis,
especially evident with attestations that additionally show apocope such as dixol (~ dixo le), diot
(~ dio te), un colpel dio (~ un golpe le dio) and quem (~ que me) (data principally from Martínez-
Gil, 2003).
44. For a fuller description of restrictions on clitic placement, see Perini (2002, Section 29.3),
which is also the source of some of these examples. Clements et al. (2011, pp. 407–411) also
presents a succinct description of enclisis and proclisis contexts for the two languages, as well as
for the contact variety Barranquenho.
© 2014. John Benjamins Publishing Company
All rights reserved
Letânia Ferreira and D. Eric Holt
(yielding a strong or weak implementation of the layering of levels of the prosodic hi-
erarchy, with the clitic in BP attaching to the phonological phrase, rather than being
incorporated into the neighboring prosodic word). Galves (n.d.) describes enclisis as a
verb-rst (V1) phenomenon in Classical Portuguese (15th-18th centures), which, in
contrast to Modern EP, shows categorical enclisis only when the verb is in absolute rst
position; otherwise variation between enclisis and proclisis obtains. She argues that
the contexts of obligatory proclisis have remained unchanged during the history of
European Portuguese, and that the pattern for EP could be formulated as “be enclitic
whenever you need (for phonological reasons) and you can (for syntactic reasons)”.
Galves and Galves (1995) adopt a thermo dynamical formalism in their study of the
change in clitic placement from Classical to Modern European Portuguese that they
argue accounts for both language acquisition and change; see also Galves, Paixão and
Britto (2005).
5.3.2 Dierences in preposition + article contractions
Likewise, there is dierential behavior in the contraction of certain prepositions with
the denite articles. While a complete treatment is well beyond the scope of the pres-
ent chapter, the extent of contractions is vastly lesser in Spanish than in GP and other
related dialects. In Spanish, the only contractions are of the masc. sg. form el with ei-
ther a (al) or de (del), while in GP, de also combines with the demonstratives of both
genders and numbers (deste, desses, daquelas) as does em (neste, nesses, naquelas). e
denite articles also combine with a, de, em and por (ao, da, nos, pelas). Old Leonese
varieties show contraction (with or without palatalization, according to the subdialect)
between the articles and some of these prepositions as well (enno, (e)no; pollo, polo),
but not de or a (del, al, as in Spanish), and also with con (conno or collo, vs. cono).45
6. Conclusion
is chapter has provided an overview of some of the major phonological dierences
between Spanish and Portuguese, discussing intermediate possibilities for outcomes of
various processes attested both in dialects of Spanish and Portuguese as well as in
other Hispano-Romance descendants like Galician and Mirandese. is broad view,
both diachronically and diatopically, is essential to a deeper understanding of these
languages and of their interrelatedness, as well as of the possibilities of language struc-
ture and evolution more generally, and we hope to have made a modest contribution
to this endeavor.
45. See Elsman & Holt (2009) for an optimality-theoretic approach to the Leonese data. ere
is much need for additional theoretical work of this broad comparative topic.
© 2014. John Benjamins Publishing Company
All rights reserved
On the partially divergent phonology of Spanish, Portuguese and points in between 
We would link to thank Leda Bisol, Dinah Callou, Laura Colantoni, José Ignacio
Hualde, Maria Helena Mateus and Mário Perini for comments and suggestions on an
early dra of this chapter. We also thank Ana M. Carvalho and Patrícia Amaral for
their editorial advice and several anonymous reviewers for their advice at dierent
stages of development.
Albano, E.C. (2001). O Gesto e suas Bordas: Esboço de fonologia acústico-articulatoria do portu-
gués brasileiro. Campinas, Brazil: Mercado de Letras.
Alves, M.M. (1999). As vogais médias em posição tônica nos nomes do português brasileiro. M.A.
Dissertation. Universidade Federal de Minas Gerais, Belo Horizonte.
Amaral, A. (1955). O dialecto caipira. AnhembiAnhembi, São Paulo: Anhembi.
Amastae, J. (1995). Variable spirantization: Constraint weighting in three dialects. Hispanic Lin-
guistics, 6(7), 265–285.
Azevedo, M.M. (1981). A Contrastive Phonology of Portuguese and English. Washington DC:
Georgetown University Press.
Azevedo, M.M. (1984). Loss of agreement in Caipira Portuguese. Hispania, 67(3), 403–409.
DOI: 10.2307/342111
Baker, G.K. (2003). Sibilant dissimilation and dispersion in the history of Spanish. Paper pre-
sented at the Linguistics Symposium on Romance Languages 33, Bloomington, IN.
Barbosa, P.A., & Albano, E. C. (2004). Brazilian Portuguese. Journal of the International Pho-
netic Association, 34(2), 227–232. DOI: 10.1017/S0025100304001756
Bayo Julve, J.C. (1998). La teoría del verso desde el punto de vista lingüístico (El sistema de versi-
cación del Cantar de Mio Cid). Ph.D. Dissertation. University of Barcelona.
Bisol, L. (1989). Vowel harmony: A variable rule in Brazilian Portuguese. Language Variation
and Change, 1, 185–198. DOI: 10.1017/S0954394500000065
Bortoni-Ricardo, S.M. (1985). e urbanization of rural dialect speakers: A sociolinguistic study
in Brazil. Cambridge: CUP.
Bouzouita, M. (2011). Future constructions in Medieval Spanish: Mesoclisis uncovered. In R.
Kempson, E. Gregoromichelaki, & C. Howes (Eds.), e dynamics of lexical interfaces (pp.
91–132). Stanford, CA: CSLI.
Bradley, T.G., & Delforge, A.M. (2006). Systemic contrast and the diachrony of Spanish sibilant
voicing. In D. Arteaga & R. Gess (Eds.), Historical Romance linguistics: Retrospectives and
perspectives (pp. 19–53). Amsterdam: John Benjamins.
Bradley, T.G. (Forthcoming). Róticas: Descripción fonética. In J. Gil & J. Llisterri (Eds.), Fo-
tica y fonología descriptivas de la lengua española. Madrid: Consejo Superior de Investiga-
ciones Cientícas.
Brandão de Carvalho, J. (1988). Nasalité et structure syllabique en portugais et en galicien: Ap-
proche non linéare et panchronique d’un problème phonologique. Verba, 15, 237–263.
Cagliari, L.C. (1997). Fonologia do português: Análise pela geometria de traços. Campinas: Edição
do Autor.
© 2014. John Benjamins Publishing Company
All rights reserved
 Letânia Ferreira and D. Eric Holt
Callou, D. & Leite, Y. (1990). Iniciação à fonética e à fonologia. Rio de Janeiro: Jorge Zahar.
Carvalho, A.M. (2006). Nominal number marking in a variety of Spanish in contact with Portu-
guese. In T. Face & C. Klee (Eds.), Selected proceedings of the 8th Hispanic Linguistics Sym-
posium, (154–166). Somerville, MA: Cascadilla Proceedings Project.
Castro, I. (2004). Introdução à história do português. Geograa da língua portuguesa antiga. Lis-
boa: Colibri.
Cintra, L.F.L. (1995). Estudos de dialectologia portuguesa. Lisboa: Sá da Costa.
Clements, G.N. (1990). e role of the sonority cycle in core syllabication. Papers in Labora-
tory Phonology, 1, 283–333. DOI: 10.1017/CBO9780511627736.017
Clements, J.C., Amaral, P., & Luís, A.R. (2011). Spanish in contact with Portuguese: e case of
Barranquenho. In M. Díaz-Campos (Ed.), e handbook of Hispanic sociolinguistics (pp.
395–417). Malden, MA: Wiley-Blackwell. DOI: 10.1002/9781444393446.ch19
Colina, S. (2009). Spanish phonology: A syllabic perspective. Washington, DC: Georgetown Uni-
versity Press.
Colina, S. (2011). Plural formation in Galician. In J. Herschensohn (Ed.), Selected proceedings
from the 40th Linguistic Symposium on Romance Languages (pp. 79–98). Amsterdam: John
Colman, F. (1983). ‘Vocalisation’ as nucleation. Studia Lingüística, 37, 30–48. DOI: 10.1111/j.1467-
Corrêa, L., Augusto, M., & Ferrari Neto, J. (2005). e early processing of number agreement in
the DP: Evidence from the acquisition of Brazilian Portuguese. In D. Bamman, T. Mag-
nitskaia, & C. Zaller (Eds.), Proceedings of the 30th Boston University Conference on Lan-
guage Development. Somerville, MA: Cascadilla Press. Retrieved from the online proceed-
ings supplement at <>; <www.les/2011/05/30-CorreaBUCLD2005.pdf>
Cristófaro, S.T. (2007). Fonética e fonologia do português (9th ed.). São Paulo: Editora Contexto.
Duell, M.J. (1999). e metric cleansing of Hispanic verse. Bulletin of Hispanic Studies, 76,
151–168. DOI: 10.3828/bhs.76.2.151
Elsman, M.M., & Holt, D.E. (2009). When small words collide: Morphological reduction and
phonological compensation in Old Leonese contractions. In R. Leow, H. Campos, & D.
Lardiere (Eds.), Little words: eir history, phonology, syntax, semantics, pragmatics, and
acquisition, (21–33). Washington, DC: Georgetown University Press.
Espírito-Santo, F. (2010). Spirantization occurrences in German and Portuguese. Máthesis, 19,
Ferrari Neto, J. (2003). Reconhecimento do número gramatical e processamento da concordância
de número no sintagma determinante na aquisição do Português Brasileiro. M. A. Disserta-
tion. Pontifícia Universidade Católica, Rio de Janeiro, RJ.
Ferreira, F. (2001). Variation in Ibero-Romance: A study of /s/ reduction in Brazilian Portuguese
in comparison with Caribbean Spanish. Ph.D. Dissertation. University of New Mexico, Al-
buquerque, NM.
Ferreira, L. (2008). High initial tones and plateaux in Spanish and Brazilian Portuguese neutral
declaratives: Consequences to the relevance of F0, duration and vowel quality as stress corre-
lates. Ph.D. Dissertation. University of Illinois, Urbana-Champaign, IL.
Ferreira, L. (2011). Peaks and plateaux: dierent F0 patterns representing the same category.
Revista Lingüística, 7(1), 17–34.
Ferreira, L., & Pérez-Leroux, A.T. (Forthcoming). Interpreting subject agreement: Data from
Brazilian Portuguese. Lingua.
© 2014. John Benjamins Publishing Company
All rights reserved
On the partially divergent phonology of Spanish, Portuguese and points in between 
Fradejas Rueda, J.M. (2000). Fonología histórica del español. Madrid: Visor Libros.
Frota, S. (2003). e phonological status of initial high peaks in European Portuguese. Catalan
Journal of linguistics, 2, 133–52.
Frota, S., & Vigário, M. (2000). Aspectos de prosódia comparada: Ritmo e entoção no PE e no
BP. In R.V. Castro & P. Barbosa (Eds.), Actas do XV encontro da associação portuguesa de
linguística (pp. 533–555). Braga: APL.
Galves, Charlotte. (n.d.). Clitic-placement in the history of Portuguese and the syntax-phonolo-
gy interface. Retrieved from <
Galves, A., & Galves, C. (1995). A case study of prosody driven language change: From Classical to
Modern European Portuguese. Retrieved from <
Galves, C., Paixão de Sousa, M.C., & Britto, H. (2005). e change in clitic placement from Clas-
sical to Modern European Portuguese: Results from the Tycho Brahe corpus. Journal of
Portuguese Linguistics, 4(1), 39–67.
Harris, J.W. (1985). Autosegmental phonology and liquid assimilation in Havana Spanish. In
L.D. King & C.A. Maley (Eds.), Selected proceedings from the 13th linguistic symposium on
Romance languages (pp. 127–148). Amsterdam: John Benjamins.
Holt, D.E. (1997). e role of the listener in the historical phonology of Spanish and Portuguese: An
optimality-theoretic account. PhD Dissertation. Georgetown University. Retrieved from the
Rutgers Optimality Archive. <les/278-0898/278-0898-HOLT-6-0.
Holt, D.E. (1998). e role of comprehension, reinterpretation and the Uniformity Condition in
historical change: e case of the development of Cl clusters from Latin to Hispano-Ro-
mance. In V. Samiian (Ed.), Proceedings of the twenty-sixth Western Conference on Linguistics,
(pp. 133–148). Fresno, CA: Department of Linguistics, California State University, Fresno.
Holt, D.E. (1999). Underspecication, constriction-based vowel geometry and scalar raising in
Asturiano. e 73rd annual meeting of the Linguistic Society of America, January 7–10, Los
Angeles, CA.
Holt, D.E. (2000). Comparative optimality-theoretic dialectology: Singular/plural nasal alterna-
tions in Galician, Mirandese (Leonese) and Spanish. In H. Campos, E. Herburger, Alfonso
Morales-Front, & T.J. Walsh (Eds.), Papers from the third Hispanic Linguistics Symposium
(pp. 125–143). Somerville, MA: Cascadilla Press.
Holt, D.E. (2002). e articulator group and liquid geometry: Implications for Spanish phonol-
ogy present and past. In C. Wiltshire & J. Camps (Eds.), Romance phonology and variation
(pp. 85–99). Amsterdam: John Benjamins.
Holt, D.E.(Ed.). (2003a). Optimality theory and language change. Dordrecht: Kluwer. DOI:
Holt, D.E. (2003b). e emergence of palatal sonorants and alternating diphthongs in Hispano-
Romance. In Holt (Ed.), 2003a (pp. 285–305).
Holt, D.E. (2004). Optimization of syllable contact in Old Spanish via the sporadic sound change
metathesis. Probus: International Journal of Latin and Romance Linguistics, 16, 43–61.
Holt, D.E. (2007). Optimality eory and language change in Spanish. In F. Martínez-Gil & S.
Colina (Eds.), Optimality-theoretic advances in Spanish phonology (pp. 378–396). Amster-
dam: John Benjamins.
Hualde, J.I. (1998). Autosegmental and metrical spreading in the vowel-harmony systems of
northwestern Spain. Linguistics, 27, 773–805.
© 2014. John Benjamins Publishing Company
All rights reserved
 Letânia Ferreira and D. Eric Holt
Hualde, J.I. (2005). e sounds of Spanish. Cambridge: CUP.
Jiménez, J., & Lloret, M.R. (2007). Andalusian vowel harmony: Weak triggers and perceptibility.
e 4th Old World Conference in Phonology (Workshop on harmony in the languages of the
Mediterranean). Rhodes, January 18–21.
Lacerda, A., & Rossi, N. (1958). Particularidades fonéticas do comportamento elocucional do
Falar do RJ. Coimbra: Impressora da Universidade de Coimbra.
Lapesa, R. (1986). Historia de la lengua española. Madrid: Gredos.
Leite de Vasconcelos, J. (1900). Estudos de philologia mirandesa. Lisboa: Imprensa Nacional.
Lipski, J.M. (1984). Latin American Spanish. London: Longman.
Lipski, J.M. (1985). Brazilian Portuguese vowel nasalization: Secondary aspects. Canadian Jour-
nal of Linguistics, 20, 59–77.
Lipski, J.M. (1989a). Spanish yeísmo and the palatal resonants: Towards a unied analysis. Pro-
bus, 1, 211–223. DOI: 10.1515/prbs.1989.1.2.211
Lipski, J.M. (1989b). /s/ voicing in Ecuadorian Spanish: Patterns and principles of consonantal
Modication. Lingua, 79, 49–71. DOI: 10.1016/0024-3841(89)90019-3
Lleó, C. (2003). Some interactions between word, foot, and syllable structure in the history of
Spanish. Studies in Natural Language and Linguistic eory, 56, 249–283. DOI: 10.1007/978-
Lloyd, P.M. (1987) From Latin to Spanish. Vol. I: Historical phonology and morphology of the
Spanish language. Philadelphia PA: e American Philosophical Society.
Lloyd, P.M. (1994). Contribution to the study of the syllable structure of Old Spanish. e Mod-
ern Language Association meeting, San Diego, California. December 28.
Luís, A., & Spencer, A. (2005). A paradigm function account of ‘mesoclisis’ in European Portu-
guese. Yearbook of Morphology, 2004, 177–228.
Manzini, M.R., & Savoia, L. M. (2011). Mesoclisis in the imperative: Phonology, morphology or
syntax? Lingua, 121,1101–1120. DOI: 10.1016/j.lingua.2011.02.002
Marins, F., & Margotti, F. (2012). Comportamento fonético-fonológico do /S/ pós-vocálico em
Manaus. Revista Investigações, 25(2), 249–274.
Martinet, A. (1974). Economía de los cambios fonéticos: Tratado de fonologia diacrónica. Madrid:
Martínez-Gil, F. (2003). Consonant intrusion in heterosyllabic consonant-liquid clusters in Old
Spanish and Old French: An optimality theoretical account. In R. Cameron, L. López & R.
Núñez-Cedeño (Eds.), A Romance perspective in language knowledge and use: Selected pa-
pers from the 31st Linguistic Symposium on Romance Languages (pp. 39–58). Amsterdam:
John Benjamins.
Martínez-Gil, F. (2007). Upstepping vowel height: a constraint-based account of metaphony in
Proto-Spanish and Lena Asturian. In F. Martínez-Gil & S. Colina (Eds.), Optimality-theo-
retic advances in Spanish phonology (pp. 99–145). Amsterdam: John Benjamins. DOI:
Masip, V. (1999). Curso de fonología y ortografía española para brasileiros. Barcelona: Difusión.
Mateus, M.H. & d’Andrade, E. (2000). e phonology of Portuguese. Oxford: OUP.
Mattoso Camara, J. (1970). Estrutura da língua portuguesa. Petrópolis: Voz e s .
Mattoso Camara, J. (1972). e Portuguese language. (A. J. Naro, Trans.). Chicago, IL: University
of Chicago Press.
McCarthy, J.J. (1984). eoretical consequences of Montañés vowel harmony. Linguistic Inquiry,
15, 291–318.
© 2014. John Benjamins Publishing Company
All rights reserved
On the partially divergent phonology of Spanish, Portuguese and points in between 
Medeiros, B. (2011). Nasal coda and vowel nasality in Brazilian Portuguese. In S.M. Alvord
(Ed.), Selected proceedings of the 5th conference on laboratory approaches to Romance pho-
nology (pp. 33–45). Somerville, MA: Cascadilla Proceedings Project.
Meillet, A. (1970). General characteristics of the Germanic languages. (W.P. Dismukes, Trans.).
Coral Gables, FL: University of Miami Press.
de Moraes Ferreira, A.J. (1898). Dialecto mirandez. Lisbon: da Silva.
Morais-Barbosa, J. (1962). Les voyelles nasales portugaises: Interprétation phonologique. In A.
Rigault (Ed.), Proceedings of the Fourth International Congress of Phonetic Sciences (pp.
691–708). e Hague: Mouton.
Morales-Front, A. & Holt, D.E. (1997). On the interplay of morphology, prosody and faithful-
ness in Portuguese pluralization. In F. Martínez-Gil & A. Morales-Front (Eds.), Issues in the
phonology and morphology of the major Iberian languages, (393–437). Washington, DC:
Georgetown University Press.
Naro, A. (1981). e social and structural dimensions of a syntactic change. Language, 57, 63–98.
Nunes, J.J. (1989). Compêndio de gramática histórica portuguesa: Fonética e morfologia (9th ed.).
Lisbon: Livraria Clássica Editora.
Oliver, Julia. (2008). Vowel raising in Puerto Rican Spanish. Ph.D. Dissertation. University of
Illinois, Chicago, IL.
Padgett, J. (1991). Stricture in feature geometry. Ph.D. Dissertation. University of Massachusetts:
(Revised version published 1995, Stanford, CA: CSLI.)
Padgett, J. (2003). e emergence of contrastive palatalization in Russian. In D.E. Holt (Ed.),
Optimality theory and language change, (307–335). Dordrecht: Kluwer. DOI: 10.1007/978-
Palácio, A. (1989). Um caso de permulta consonantica no dialeto do Recife. In A.T. Castilho
(Ed.), Português falado no Brasil, (25–33). Campinas: Editora UNICAMP.
Parkinson, S. (1983). Portuguese nasal vowels as phonological diphthongs. Lingua, 61, 157–77.
DOI: 10.1016/0024-3841(83)90031-1
Parkinson, S. (1988). Portuguese. In M. Harris & N. Vincent (Eds.), e Romance languages (pp.
131–169). Oxford: OUP.
Peixoto, J. (2011). O ditongo em Português: História, variação e gramática. Revista Linguística,
7, 77–111.
Penny, R. (1969). Vowel harmony in the speech of the Montes de Pas (Santander). Orbis, 18,
Penny, R. (1991). A history of the Spanish language. Cambridge: CUP.
Penny, R. (2009). Vowel harmony and metaphony in Iberia: A revised analysis. Estudos de
Lingüística Galega, 1, 113–124.
Perez, J.I. (1982). Observaciones en torno a la desaparición de la -N- intervocálica en gallego.
Verba, 9, 201–213.
Perini, M.A. (2002). Modern Portuguese: A reference grammar. New Haven, CT: Yale University
Porzio Gernia, M L. (1976). Tendenze strutturali della sillaba latina in età arcaica e classica.
Scritti in onore de G. Bonfante (pp. 757–779). Brescia: Paideira Editrice.
Quilis, A. (1993). Tratado de fonología e fonética españolas. Madrid: Editorial Gredos.
Reed, D., & Leite, Y. (1956). e segmental phonemes of Brazilian Portuguese: Standard Pau-
lista dialect. In K. Pike (Ed.), Phonemics: A technique for reducing language to writing (pp.
194–202). Ann Harbor, MI: University of Michigan.
Sampson, R. (1999). Nasal vowel evolution in Romance. Oxford: OUP.
© 2014. John Benjamins Publishing Company
All rights reserved
 Letânia Ferreira and D. Eric Holt
Scherre, M. (2001). Phrase-level parallelism eect on noun phrase number agreement. Lan-
guage Variation and Change, 13, 91–107. DOI: 10.1017/S0954394501131042
Scherre, M., & Naro, A. (2006). Mudança sem mudança: A concordância de número no português
brasileiro. Scripta, 9, 107–129.
Selkirk, E. (1984). On the major class features and syllable theory. In M. Arono & R. Oehrle
(Eds.), Language sound structure: Studies in phonology (pp. 107–136). Cambridge, MA: e
MIT Press.
Shosted, R. (2006). Vocalic context as a condition for nasal coda emergence: Aerodynamic evi-
dence. Journal of the International Phonetic Association, 36,39–58. DOI: 10.1017/
Simioni, T. (2008). e clitic and its place in the prosodic structure in Brazilian Portuguese. Alfa,
52(2), 431–446.
Vásquez Cuesta, P., & Mendes da Luz, M.A. (1980). Gramática da língua portuguesa. Lisboa:
Edições 70.
Veloso, J. (2002). Do fricative+plosive onsets exist word-initially in European Portuguese? In K.
Petrova, A. Eimova, & R. Shopov (Eds.), Litora psycholinguistica (pp. 49–56). Sóa: Sema
Vieira, M.J. (2002). As vogais médias postônicas: Uma análise variacionista. In L. Bisol & C.
Brescancini (Eds.), Fonologia e variação: Recortes do Português, (127–159). Porto Alegre:
Walker, R. (2005). Weak triggers in vowel harmony. Natural Language and Linguistic eory, 23,
917–989. DOI: 10.1007/s11049-004-4562-z
Walker, R. (2011). Vowel patterns in language. Cambridge: CUP. DOI: 10.1017/CBO97805119
Walsh-Dickey, L. (1997). e phonology of liquids. Ph.D. Dissertation. University of Massachu-
setts, Amherst, MA.
Wetzels, W.L. (1991). Contrastive and allophonic properties of Brazilian Portuguese vowels. In
D. Wanner & D.A. Kibbee (Eds.), New analyses in Romance linguistics (pp. 77–99). Amster-
dam: John Benjamins.
Wetzels, W.L. (1995). Midvowel alternations in the Brazilian Portuguese verb. Phonology, 12,
281–304. DOI: 10.1017/S0952675700002505
Williams, E.B. (1962). From Latin to Portuguese: Historical phonology and morphology of the
Portuguese language. Philadelphia PA: University of Pennsylvania Press.
Wireback, K.J. (1996). On the palatalization of /kl/, /pl/, and // in Hispano-Romance. Paper pre-
sented at the 49th Kentucky Foreign Language Conference, University of Kentucky, April 19.
Zec, D. (1995). Sonority constraints on syllable structure. Phonology, 12, 85–129. DOI: 10.1017/
Zubizarreta, M.L. (1979). Vowel harmony in Andalusian Spanish. MIT Working Papers in Lin-
guistics, 1,1–11.
... Cano González, 2009, p. 90;García Arias, 1974, p. 158) en (3), muestra las principales características de las diferentes categorías prosódicas (cfr. Blevins, 1995;Calabrese, 2005;Ferreira y Holt, 2014). La sílaba (σ) se representa mediante una estructura jerárquica con los siguientes constituyentes: el ataque (A), la rima (R), el núcleo (N) y la coda (C). 4 El núcleo representa el centro de una sílaba, mientras que el ataque y la coda constituyen sus extremos. ...
Full-text available
Na nuesa contribución presentaremos un estudiu sobro l’adautación de les consonantes nos empréstamos n’asturianu. Fadremos por analizar el comportamientu fonolóxicu de les consonantes de los empréstamos dende un puntu de vista silábicu. Naguamos por describir más precisamente’l so grau d’adautación en rellación a les restricciones prosódiques que rixen la cadarma silábica d’esta llingua y facer comparanza de los resultaos con otros idiomes románicos. El casu del asturianu dexarános reflexonar tamién sobro los problemes y les cuestiones que se planteguen davezu na lliteratura dedicada a l’adautación de los empréstamos: ¿cómo se produz l’adautación d’un empréstamu nuna llingua? ¿Hai preferencies universales pa dellos tipos de reparación sobro otros, por casu la epéntesis sobro la supresión? ¿Han tratase los empréstamos de mou uniforme dientro’l léxicu o tán suxetos a parámetros especiales acordies col so grau d’inxerimientu? Pallabres clave: empréstamos, morfo-fonoloxía, sílaba, restricciones.
... This strong correlation between central and back vowels and /s/ aspiration is also linked to a monophthongization phenomenon that is very common in Brazilian Portuguese, especially in the Rio de Janeiro dialect. As observed in the data, the palatal glide [j] is often inserted before /s/ in some wordfinal environments (e.g., nós 'we' [nɔjs], vez 'turn' [veis] as in 'my turn') as a result of epenthesis (Ferreira & Holt 2014;Perini 2002). So, to create an environment where /s/ aspiration can occur, i.e., not preceded by high front vowels, the palatal glide [j] is often deleted from falling diphthongs in syllable-coda position. ...
Full-text available
Following the theoretical and methodological principles of Variationist Sociolinguistics, this paper analyzes the use of the aspirated variant of postvocalic /s/ by residents of City of God, a predominantly-black neighborhood in Rio de Janeiro that is widely known as a favela (roughly, shantytown or slum). The analyzed data consist of seventeen sociolinguistic interviews conducted in 2015 with twenty-two residents of this community. The quantitative analysis included six social variables—race/color, regional origin, age, gender, education, and speaker—as well as six linguistic variables—preceding vowel, following sound, syllabic stress, number of syllables, grammatical category, and word. Race/color, age, and all the linguistic factors considered in the analysis were selected as statistically significant to the occurrence of /s/ aspiration. This study indicates a possible connection between aspiration and race/color and stresses the importance of including racial identity as a relevant factor in sociolinguistic studies in Brazil, especially those focusing on favelas and other similar urban communities.
... If we extend the phenomena of obstruent lenition due to loss of length distinction to /l lː n nː/, then it is possible that the simplification of geminate /lː nː/ to singleton /l n/ was a related process. Indeed, Holt (1997;2016) and Ferreira and Holt (2014) postulate that /lː nː/ were able to simplify without causing confusion, which implies that geminate simplification in Galician-Portuguese was a later development than the loss of intervocalic /l n/. ...
Full-text available
From Latin to Spanish and Portuguese, the natural class of sonorant consonants – laterals, rhotics, and nasals – often underwent drastic phonological changes. It is noteworthy that the tendency toward dissimilation, in accordance with Dispersion Theory (Flemming 1996; 1997; 2006), effected opposite changes in intervocalic /l/ and /n/. Portuguese favored geminate simplification and singleton lenition (Lief 2006; Malkiel & Alessandri Teixeira 1985), whereas Spanish tended toward geminate palatalization and singleton retention (Lloyd 1987). This study is an expansion of Holt (2007) and presents a diachronic and contrastive analysis of the evolution of intervocalic /l lː n nː/ from Latin to Spanish and Portuguese.
Full-text available
Optimality Theory and Language Change: -discusses many optimization and linguistic issues in great detail; -treats the history of a variety of languages, including English, French, Germanic, Galician/Portuguese, Latin, Russian, and Spanish; -shows that the application of OT allows for innovative and improved analyses; -allows researchers that appeal to OT to see the connections of their (usually synchronic) work with diachronic studies; -contains a complete bibliography on Optimality Theory and language change. This volume may be used as one of the texts in courses on historical phonology or syntax that treat these topics from generative approaches or that give a general survey of various frameworks of research into these areas. Likewise, the volume may serve as a text for courses in phonology, syntax and Optimality Theory that have a component dedicated to extensions of linguistic theory to historical change. It is of interest for historical linguists, researchers into Optimality Theory and linguistic theory, and for phonologists and syntacticians with an interest in historical change.
Full-text available
This thesis examines the interplay of fundamental frequency, duration and vowel quality as acoustic correlates of stress in Brazilian Portuguese and Peninsular Spanish. More specifically this study investigates the role of each one of these correlates in determining stress in these languages and elaborates on how the other correlates behave in dealing with the limitations or complexity of information provided by the other correlates when expressing stress. The working hypotheses of this study are as follows.1) Portuguese presents a very complex intonational structure; as a result, F0 functions as a secondary acoustic correlate of stress in Brazilian Portuguese. 2) Spanish presents a simpler intonational structure than Brazilian Portuguese, and intonation in Spanish is extremely reliable when expressing lexical stress. 3) Duration is an important acoustic correlate of stress for both Portuguese and Spanish, but the function of duration in expressing Portuguese stress is very complex and crucial, unlike in Spanish in which this task is much simpler. 4) Stress will affect vowel quality in both languages studied here, however, as the quality of vowels undergoes much stronger changes in Portuguese than Spanish, the role of vowel quality in expressing stress in this first language is more powerful than in the last one. In order to verify these hypothesis a direct comparison containing neutral declaratives in both Spanish and Brazilian Portuguese was elaborated. The results show that indeed Portuguese presents a more complex intonational system than Spanish in declaratives sentence initial position. Results suggest that pitch in Brazilian Portuguese sentence initial position not only express lexical stress, but phrasal tones as well. According to these results the phrasal tones in Brazilian Portuguese can be presented in 2 the intonational contour as either peaks or plateaux. Results also suggest that duration in Spanish does not play a crucial role as it does in Portuguese. Additionally, the outcome of this research indicates that Brazilian Portuguese has developed a very acute and singular system to express stress trough duration. This system seems to be adjusted according to segmental quality and to apply to different dialects over the Brazilian country. This research also shows that the effect of stress in vowel quality is much more consistent in Brazilian Portuguese than in Spanish. Together, these results have implications in the way in which these three acoustic correlate interplay in order to express stress in these two languages.
Full-text available
This paper revisits results of previous studies about sentence initial high tones in the Portuguese language and argues that the shape and distribution of these sentence initial pitch events in Brazilian Portuguese are more complex than previously suggested in the literature. By pointing out that peaks and plateaux are manifestations of the same pitch event the results also suggest that the peak, the highest point of a pitch excursion, is not always the main pitch target of the phra-sal tone in this language. These ndings contribute to the phonology-phonetics mapping in intonation.
Linguists researching the sounds of languages do not just study lists of sounds but seek to discover generalizations about sound patterns by grouping them into categories. They study the common properties of each category and identify what distinguishes one category from another. Vowel patterns, for instance, are analysed and compared across languages to identify phonological similarities and differences. This account of vowel patterns in language brings a wealth of cross-linguistic material to the study of vowel systems and offers theoretical insights. Informed by research in speech perception and production, it addresses the fundamental question of how the relative prominence of word position influences vowel processes and distributions. The book combines a cross-linguistic focus with detailed case studies. Descriptions and analyses are provided for vowel patterns in over 25 languages from around the world, with particular emphasis on minor Romance languages and on the diachronic development of the German umlaut.