No full-text available
To read the full-text of this research,
you can request a copy directly from the author.
Dissimilarity in the Arabic Verbal Roots
... In phonology, distance measures inform the formulation of constraints on alternations (Gildea & Jurafsky 1996) and phonotactics (e.g. Frisch et al. 1997, Pierrehumbert 1993, and specifically in phonotactics, phonological distance is a crucial factor in modeling phonological neighborhood density, the degree to which a sound sequence overlaps with existing words in the lexicon. Models built on phonological distance measures have been applied to spoken word recognition as a predictor in experimental paradigms (Luce et al. 2000, Luce & Pisoni 1998, to the investigation of speech errors (Vitevitch 1997), and to the explanation of some phonological phenomena such as asymmetries between roots and affixes (Ussishkin & Wedel 2002). ...
... The number of different feature values is normalized by dividing by the total number of phonological features, as shown in 1. 2 This method is called HAMMING DISTANCE measure (e.g. Gildea & Jurafsky 1996, Pierrehumbert 1993). ...
... 2 Null features are usually thought to be different from both positive and negative values (Pierrehumbert 1993). We will adopt this assumption in this study, except when using Broe's information gain weighting (see Appendix A). ...
... Such dissimilatory tendencies have been reported for a variety of languages. For instance, Arabic has co-occurrence restrictions that apply to all consonant places and manners (Pierrehumbert 1993;Frisch et al. 2004), even though the restrictions are less strong for coronals (Coetzee and Pater 2008). Japanese rendaku limits voiced obstruents within a word, and while this restriction is nearly categorical in native Yamato words, speakers extend the tendency productively to novel words in a gradient fashion, with certain combinations being over-or under-represented (Kawahara and Sano 2014a,b). ...
... Second, co-occurrence restrictions are often sensitive to the identity between the interacting segments. There are at least two cross-linguistic tendencies: one kind shows fewer co-occurrence restrictions when the segments are identical (Gallagher and Coon 2009;Gallagher 2010b), whereas the other shows the opposite, namely stronger co-occurrence restrictions when segments are identical (Vance 1991;Pierrehumbert 1993). Our experiments confirm that Slovenian is of the second kind: identical postalveolars lead to lower acceptability ratings. ...
... We found that multiple velars do not have the same negative effect on acceptability as multiple postalveolars do, even though both types of co-occurrences are underrepresented in the lexicon. This contrasts with the patterns observed in Arabic, among many other languages, where co-occurrence restrictions apply across all kinds of consonants (Pierrehumbert 1993;Frisch et al. 2004). ...
This paper shows that a postalveolar co-occurrence restriction (Obligatory Contour Principle, OCP) is a productive component of Slovenian phonology. We first examine whether an apparent OCP-based restriction on derived palatalization, previously observed in corpus data (Jurgec 2016), extends to novel forms via a goodness-rating task. We then explore the generality of the restriction across the lexicon, in non-derived novel words as well as derived forms. Our results confirm that native speakers judge derived palatalized nonce forms to be less acceptable when the stem contains another postalveolar, reflecting the pattern found in the previous corpus study. We further demonstrate that multiple postalveolars are dispreferred even in non-derived words, which suggests that the effect is a general case of OCP. This is additionally supported by effects of proximity (the restriction is stronger for postalveolars separated only by a single vowel than for those further apart from one another) and identity (the restriction is stronger for identical than non-identical postalveolars), reflecting cross-linguistic tendencies in the manifestation of OCP and non-local consonant dissimilation. Finally, we show that the restriction does not appear to apply to all places of articulation, suggesting that the co-occurrence restriction in Slovenian specifically targets postalveolars, and adding a previously unattested pattern to the typology of OCP phenomena on consonant place.
... Hence, segmental OCP-Place restricts the co-occurrence of homorganic segments within a root (McCarthy 1986(McCarthy , 1988(McCarthy , 1994, the strongest restriction being among labial segments, followed by dorsals and then coronals. Additionally, Pierrehumbert (1993) and Frisch et al. (2004) argue that in addition to the major place feature, manner and voicing have effects on the co-occurrence restrictions of different segments in Arabic. They also propose that OCP-Place in Arabic is gradient, rather than absolute: the more features shared between a pair of consonants, the stronger the restriction on their co-occurrence. ...
... However, with respect to the items in (9) and (10), the co-occurrence of [Ã] and [d] is more restricted when these segments are strictly adjacent, as in (9), but not as much when they are separated by a vowel, as in (10). That is, proximity plays a role in OCP-Place restrictions on phonological alternations in Arabic, as is the case for static patterns (for static patterns, see Pierrehumbert 1993, Frisch et al. 2004. Since the OCP-Place constraint given in (4) above, repeated here as (12), is applicable to sequences with an intervening vowel, a more restrictive constraint is needed to rule out structures in which [Ã] and [d] are strictly adjacent. ...
... Specifically, some features contribute more to the identity and the perception of that identity of a given segment. This observation is also made by Pierrehumbert (1993) and Frisch et al. (2004), who refer to the perceived similarity 14 The vertical line between *([anterior] ∼ [anterior]) coronal and *([anterior] . . . ...
This paper provides evidence for the activity of the Obligatory Contour Principle (OCP) as a constraint on dynamic alternations in the synchronic grammar of Qatari Arabic. It shows that the OCP is subject to proximity and to a gradient similarity effect. In Qatari Arabic, there are two variable phonological alternations that interact with the OCP, affrication and lenition. The velar stops /g/ and /k/ affricate to [ʤ] and [ʧ], respectively, when adjacent to [i(:)]. However, affrication is blocked when the outcome includes a sequence of segments that are highly similar. Lenition applies variably to the phoneme /ʤ/ , which surfaces as [ʤ] or [j]. Usually, the probability of lenition applying to its eligible candidates is around the level of chance. The process, however, applies categorically when a violation of the OCP would otherwise be incurred. The data are analyzed within the framework of Optimality Theory.
... In the preceding section we saw how the OCP operates at the segmental level (root tier) to prohibit adjacent identical consonants. This can be termed the 'total OCP' (Pierrehumbert 1993), since it refers to adjacent consonants that are identical in all their features. But the OCP is also important in explaining another widespread generalization regarding Semitic roots. ...
... This approach has the weakness that it is categorical: it predicts that /gfb/ and /fgb/ should be equally disfavored, since the Labial nodes that constitute the OCP violation are equally adjacent in each case. However, as Pierrehumbert (1993) shows for Arabic, the degree to which coocurrence of homorganic consonants is disfavored correlates with their proximity in the root: /gfb/ is worse than /fgb/. Further, the degree of disfavoring is proportional to the relative identity of the homorganic consonants in features beyond those of place of articulation. ...
... The basic approach taken here, as in Pierrehumbert (1993), is to compare the expected cooccurrences of each pair of consonants with the occurrences which are actually attested. The method here differs since three types of roots are included, rather than just triliterals. ...
... All currently popular theories of grammar subscribe to the cognitive commitment: we are interested in describing what speakers of human languages know about the languages they speak, and the sound patterns of those languages in particular, not simply in describing the corpus of utterances we observe in the most parsimonious way possible (Albright & Hayes 2003, Bybee 2001, Chomsky & Halle 1968, Daelemans & van den Bosch 2005, Goldberg 1995, Langacker 1987, Nesset 2008, Prince & Smolensky 2004[1993). All theories that make a cognitive commitment place restrictions on the types of generalizations language learners make on the basis of the primary linguistic data they are exposed to. ...
... 3 Massaro (1970) likewise shows the fallibility of sound-comparison processes: whether two stimuli are acoustically identical is difficult to determine if they are not temporally adjacent. Pierrehumbert (1993) uses the proposal that between-segment comparisons are fallible to account for the distance effect on the obligatory contour principle. ories arising from the functional/cognitive/constructionist tradition, is what Bybee (1985, Bybee & Moder 1983, Bybee & Slobin 1982 calls PRODUCT-ORIENTED SCHEMAS, also known as CONSTRUCTIONS (Booij 2008, Goldberg 1995 and FIRST-ORDER SCHEMAS (Langacker 1987, Nesset 2005. ...
... Second, product-oriented generalizations are supported by examples of rule conspiracies (Kisseberth 1970), in which a diverse collection of changes results in avoiding or producing the same sound sequence. The importance of capturing such patterns has been recognized in phonology and triggered a paradigm shift from rule-based phonology to OPTIMALITY THEORY (OT; Kager 1999, Prince & Smolensky 2004[1993), which can capture some product-oriented generalizations using markedness constraints. ...
This article reports on an experiment with miniature artificial languages that provides support for a synthesis of ideas from USAGE-BASED PHONOLOGY (Bybee 1985, 2001, Nesset 2008) and HARMONIC GRAMMAR (Legendre et al. 1990, Smolensky & Legendre 2006). All miniature artificial languages presented to subjects feature velar palatalization (k → tſ) before a plural suffix, -i. I show that (i) examples of -i simply attaching to a [tſ]-final stem help palatalization (supporting t → tſi over t → ti and p → tſi over p → pi), a finding that provides specific support for product-oriented schemas like 'plurals should end in [tſi]'; (ii) learners tend to perseverate on the form they know, leveling stem changes, which provides support for paradigm-uniformity constraints in favor of retaining gestures composing the known form, for example, 'keep labiality'; and (iii) the same plural schema helps untrained singular-plural mappings more than it helps trained ones. This result is accounted for by proposing that schemas and paradigm-uniformity constraints clamor for candidate plural forms that obey them. Given that competition is between candidate outputs, the same schema provides more help to candidates that violate strong paradigm-uniformity constraints and are therefore weak relative to competitor candidates. A computational model of schema extraction is proposed.
... In this constraint, originally proposed as an instantiation of the Obligatory Contour Principle (Goldsmith 1979), repeated place of articulation features are not allowed within a root. Subsequent research has shown that the details of consonant occurrence in the Arabic roots are complex, with the strength of the phonotactic restriction gradiently dependent on the similarity of the consonants involved, the presence of intervening segments, and the contrasts available in the segmental inventory of Arabic (Pierrehumbert 1993;Frisch, Pierrehumbert, and Broe in press). The gradience of the phonotactic patterns in the Arabic lexicon provide strong evidence for a functional phonetic motivation for the constraint. ...
... Analogous constraints that apply to repeated laryngeal features rather than repeated place features are also attested across unrelated languages such as Sanskrit, Hausa, and Souletin Basque (MacEachern 1999). Further, in cases where lexical patterns have been analysed statistically, the co-occurrence patterns are gradient and quantitatively depend on similarity (Berkley 1995(Berkley , 2000Buckley 1997;Frisch 1996;Pierrehumbert 1993). ...
... When the segmental OCP constraints within a particular language are examined more closely it becomes apparent that homorganic segments are constrained to different degrees, suggesting that the all-or-nothing autosegmental OCP cannot properly explain the patterns. Pierrehumbert (1993) observed that the Arabic consonant co-occurrence constraints depend on the similarity of the homorganic consonant pairs involved. For example, the alveolar consonants divide into distinct series of obstruents and sonorants, where co-occurrence between the series is frequent, while co-occurrence within the series is uncommon (Greenberg 1950;McCarthy 1994). ...
Introduction The Arabic verbal roots are subject to a long-distance phonotactic constraint that is well known for its implications for autosegmental representation (McCarthy 1986, 1988, 1994). In this constraint, originally proposed as an instantiation of the Obligatory Contour Principle (Goldsmith 1979), repeated place of articulation features are not allowed within a root. Subsequent research has shown that the details of consonant occurrence in the Arabic roots are complex, with the strength of the phonotactic restriction gradiently dependent on the similarity of the consonants involved, the presence of intervening segments, and the contrasts available in the segmental inventory of Arabic (Pierrehumbert 1993; Frisch, Pierrehumbert, and Broe 2004). The gradience of the phonotactic patterns in the Arabic lexicon provide strong evidence for a functional phonetic motivation for the constraint. The similarity avoidance constraint in Arabic is quantitatively dependent on similarity, distance between segments, segment frequency, and segmental position in the word. No formal model that prohibits feature co-occurrence like the autosegmental OCP can capture the richness of the patterning. However, a wide range of evidence from psycholinguistics suggests that processing a sequence of similar items is more difficult than processing a sequence of dissimilar items. Thus, we can account for the presence of similarity avoidance constraints in the phonotactics of Arabic as a consequence of functional pressure to make language processing as easy as possible. I claim that the richness of phonotactic patterns directly (quantitatively) reflects the functional explanation. In this way, statistical analysis of the lexicon provides a novel type of evidence for functionally motivated constraints and rules out alternative formal explanations (see Hawkins 1994 for similar arguments at the syntactic level).
... OCP effects) among consonants (e.g. Pierrehumbert 1993, Frisch et al. 2004, though there has also been similar work on assimilatory patterns (e.g. Rose & King 2007, Brown 2008, Arsenault 2012. However, most of this work has focused primarily on consonant co-occurrences. ...
... The typical measure used to quantify gradient cooccurrence restrictions is observed over expected (O/E) values (e.g. Pierrehumbert 1993;Frisch et al. 2004;Coetzee & Pater 2008 Beyond O/E, harmony has also been quantified overall for a language (Harrison et al. 2004). This value expresses the extent to which the percentage of harmonic co-occurrences in a corpus exceeds chance, given the relative proportions of each harmonic category in the corpus. ...
Despite substantial phonological research into segmental co-occurrence patterns, there is currently no systematic way of calculating the gradient degree to which a segment participates in a harmony system, across its co-occurrences with all other segments. In this paper, I adopt the statistical concept of relative risk as a measure of participation in harmony. I compute both O/E values and the relative risk measure for vowels in corpora of three languages with front/back harmony: Chuvash, Tatar, and Mari. I show that relative risk corresponds to the intuitive notion of how much a vowel participates in harmony, viewed based on how regularly it occurs in disharmonic contexts. I then consider the implications of the results, given what is known about categorical trends of participation in front/back harmony systems in other languages. For example, the relative risk values show that [i] generally participates less in the harmony system than most other vowels in all of these languages, and that marked vowels are typically highly harmonic. As such, this measure can illuminate gradient language-internal and cross-linguistic patterns in harmony participation that are not apparent from more categorical descriptions or entirely clear from O/E values.
... Older studies include Uhlenbeck 1949, Chastaing 1964. More recently, Pierrehumbert 1993, Frisch 1996, Frisch et al. 1997, and Frisch & Zawaydeh 2001 investigated statistical patterns of the Arabic lexicon (in particular, the effects of consonant coocurrence constraints in the verbal roots of Arabic). Pierrehumbert 1994 deals with dissimilarity requirements operating on English consonant clusters across intervening material. ...
... These constraints on contrasts can be divided into two distinct types with opposite effects: constraints that preserve distinctness and dissimilarity, and thus maximize formal contrasts, and constraints enforcing structural simplicity, thus minimizing contrast. Illustrations of the former are the constraints on (highly) similar homorganic consonant pairs: such pairs are found less frequently than dissimilar homorganic consonant pairs (the OCP effect; Pierrehumbert 1993, Frisch et al. 1997). An instance of such a constraint is the Dutch constraint on identical homorganic liquid consonants (see §3.4). ...
... As classes de afixos podem, por sua vez, ser identificadas por restrições fonotáticas. Pesquisas recentes (Pierrehumbert 1993, Frisch 1996 mostram, além disso, que algumas dessas restrições são melhor entendidas não como categóricas e atreladas a ambientes fixos, mas como probabilísticas e direcionais. ...
... A metodologia de análise baseia-se numa sugestão de Greenberg (1950), retomada, desenvolvida e refinada por Pierrehumbert (1993Pierrehumbert ( , 1994 e Frisch (1996), segundo a qual o léxico pode ser encarado como um conjunto de distribuições nãoaleatórias de probabilidades não só de ocorrência mas também de combinação de unidades fônicas. Daí decorre que a gramática fônica não é mais do que o conjunto desses vieses lexicais, que exprimem ora a atração, ora a repulsão da língua por certas unidades ou conjuntos delas. ...
This paper contends that the two competing "rules" that the literature on Portuguese morphophonology has claimed to apply to the verb paradigm, namely, vowel height harmony and vowel lowering, are, in fact, phonotactic restrictions that apply, in a categorical fashion, to the inflected verb stem and, in a gradient fashion, to the non-inflected verb stem. At least in Brazilian Portuguese, the non-inflected verb stem is consistent with the inflected verb stem in that lowering predominates in both in the first conjugation and harmony predominates in both in the second and third conjugation. Lowering is in turn consistent with other versions of OCP which cut across all grammatical categories. The findings are interpreted in light of Acoustic-Articulatory Phonology (Albano 2001), which predicts, on grounds of facilitation of decoding of acoustic-to-articulatory relations, that stress tends to attract low vowels except where vowel quality is otherwise predictable.
... The empirical findings from the current study are also in line with a more general tendency that the strength of phonotactics decays as the distance of phonotactics increases as found in many studies (Frisch 1996;Hayes & Zsuzsa 2006;Kharitonov 1982;Kimper 2011;Pierrehumbert 1993;Zymet 2014). It has been found in various studies using different approaches that the strictness of phonotactic strengths decreases as the distance of two phonemes increases. ...
An underlying assumption in terms of sonority distances is that clusters with large sonority distances are more common than those with small distances, as captured in the unmarked status of large sonority distances and formalized in terms of sonority constraints on consonant clusters. A cross-linguistic survey of attested sonority distances in 357 languages reveals that large sonority distances are not most commonly attested. Rather, there is a point of sonority distance at which the largest number of languages is attested. When the sonority distance exceeds a particular value, the number of languages starts to decrease, regardless of the sonority scales tested. The finding puts the unmarked status of large sonority distances to the test, suggesting a potential constraint that prevents large distances from surfacing.
... The Frequency-Driven Constraint Induction mechanism of STAGE calculates observed/expected ratios (O/E; Pierrehumbert, 1993;Frisch et al., 2004) of all biphones that occur in the input data and induces constraints by setting thresholds on the O/E ratios. O/E ratios compare how often a biphone actually occurs in the data (Observed) to how often each biphone should have occurred if all segments are assumed to have an equal likelihood of combining to form biphones (Expected) by dividing the probability of a biphone (xy) divided by the product of the summed probability of all biphones beginning with (x) and the summed probability of all biphones ending with (y). ...
Illusory epenthesis is a phenomenon in which listeners report hearing a vowel between a phonotactically illegal consonant cluster, even in the complete absence of vocalic cues. The present study uses Japanese as a test case and investigates the respective roles of three mechanisms that have been claimed to drive the choice of epenthetic vowel—phonetic minimality, phonotactic predictability, and phonological alternations—and propose that they share the same rational goal of searching for the vowel that minimally alters the original speech signal. Additionally, crucial assumptions regarding phonological knowledge held by previous studies are tested in a series of corpus analyses using the Corpus of Spontaneous Japanese. Results show that all three mechanisms can only partially account for epenthesis patterns observed in language users, and the study concludes by discussing possible ways in which the mechanisms might be integrated.
... [3] Several studies have shown that identity aversion increases with proximitythe closer the identical elements, the stronger their aversion (Pierrehumbert 1993, Suzuki 1998, Frisch, Pierrehumbert & Broe 2004, Walter 2007. Here, however, we only examine whether people ban adjacent identical elements (e.g. in forms like panana); whether non-adjacent identical elements are further disliked (e.g. ...
Does knowledge of language transfer spontaneously across language modalities? For example, do English speakers, who have had no command of a sign language, spontaneously project grammatical constraints from English to linguistic signs? Here, we address this question by examining the constraints on doubling. We first demonstrate that doubling (e.g. panana ; generally: ABB) is amenable to two conflicting parses (identity vs. reduplication), depending on the level of analysis (phonology vs. morphology). We next show that speakers with no command of a sign language spontaneously project these two parses to novel ABB signs in American Sign Language. Moreover, the chosen parse (for signs) is constrained by the morphology of spoken language. Hebrew speakers can project the morphological parse when doubling indicates diminution, but English speakers only do so when doubling indicates plurality, in line with the distinct morphological properties of their spoken languages. These observations suggest that doubling in speech and signs is constrained by a common set of linguistic principles that are algebraic, amodal and abstract.
... (Clements & Keyser, 1983;Davis, 1989;Fudge, 1987). For another example, in Arabic triconsonantal roots, in addition to constraints against identity of the first two consonants (Greenberg, 1950;McCarthy, 1986), Pierrehumbert (1993) finds more general statistical biases against co-occurrence of consonantal features. Such statistical biases could arise from lexicalization of interactions between gestures which are not simultaneously active. ...
Most analyses of articulatory processes in speech assume that word form-related changes in the state of the vocal tract have well-defined beginnings and ends. But how do we determine the precise moments in time when these beginnings and ends occur? More specifically, when should we expect information related to the sound categories of a word to be present in acoustic and articulatory signals? The framework of Articulatory Phonology/Task Dynamics predicts that the earliest time such information becomes available is when the first articulatory gesture of a word becomes active, which closely corresponds to when a movement is initiated. Alternatively, a recent extension of the Articulatory Phonology model holds that gestures may have an influence on the state of the vocal tract after they have been retrieved from memory, but before they become active and before canonical movement initiation. This paper presents evidence that indeed, anticipatory information is available much earlier than is typically assumed: the identity of a syllable onset gesture can be predicted from articulatory and acoustic data quite early, in some cases nearly half a second before movement initiation. Likewise, the identity of a coda gesture can be predicted during the period of time typically associated with an onset consonant. These findings were obtained with a novel analysis method called signal chopping which was paired with deep neural network based classification. In this approach articulatory and acoustic signals are systematically truncated in space and time, and a network training/test procedure is repeated on the chopped signals. By analyzing the effects of chopping on classification accuracy, gesture-specific information can be spatiotemporally localized.
... The discussion in this appendix is based on the parallel example given by Wilson and Obdeyn (2009), who address the well-known observed/expected statistic (Pierrehumbert 1993) of which Martin's shuffling system is a variant; see also Jurafsky & Martin 2019:Ch. 5 for discussion of the ability of MaxEnt to disentangle correlated factors. ...
... (Clements & Keyser, 1983;Davis, 1989;Fudge, 1987). For another example, in Arabic triconsonantal roots, in addition to constraints against identity of the first two consonants (Greenberg, 1950;McCarthy, 1986), Pierrehumbert (1993) finds more general statistical biases against co-occurrence of consonantal features. Such statistical biases could arise from lexicalization of interactions between gestures which are not simultaneously active. ...
Most analyses of articulatory processes in speech assume that word form-related changes in the state of the vocal tract have well-defined beginnings and ends. But how do we determine the precise moments in time when these beginnings and ends occur? More specifically, when should we expect information related to the sound categories of a word to be present in acoustic and articulatory signals? The framework of Articulatory Phonology / Task Dynamics predicts that the earliest time such information becomes available is when the first articulatory gesture of a word becomes active, which closely corresponds to when a movement is initiated. Alternatively, a recent extension of the Articulatory Phonology model holds that gestures may have an influence on the state of the vocal tract after they have been retrieved from memory, but before they become active and before canonical movement initiation. This paper presents evidence that indeed, anticipatory information is available much earlier than is typically assumed: the identity of a syllable onset gesture can be predicted from articulatory and acoustic data quite early, in some cases nearly half a second before movement initiation. Likewise, the identity of a coda gesture can be predicted during the period of time typically associated with an onset consonant. These findings were obtained with a novel analysis method called signal chopping which was paired with deep neural network based classification. In this approach articulatory and acoustic signals are systematically truncated in space and time, and a network training/test procedure is repeated on the chopped signals. By analyzing the effects of chopping on classification accuracy, gesture-specific information can be spatiotemporally localized.
... While originally defined with reference to adjacent phonological elements (on some level of representation), later studies have shown that the OCP holds also for non-adjacent elements, though the closer the elements, the stronger the effect(Pierrehumbert 1993, Rose 2000, Frisch et al. 2004, Yeverechyahu 2014. ...
... The latter finding reflects the quasi-universal tendency that the more similar two consonants are, the more strongly a sequence of them is disfavored (e.g. Berent and Shimron 2003;Berent et al. 2004;Buckley 1997;Frisch et al. 2004;Greenberg 1950;Pierrehumbert 1993). In Experiment II, I will test whether the OCP-labial effect is active in nicknaming processes as well as rendaku. ...
The current study deals with two topics. One is the new nicknaming trend in Japanese whereby [h] alternates with [p]. In Experiment I, I established the hypothesis that the process is driven to express cuteness, and experimentally demonstrated that singleton [p] is more likely to be associated with cuteness than other consonants in Japanese. The other topic discussed in the current paper is the orthographic Lyman’s Law, or OCP(diacritic) (Kawahara, Shigeto. 2018. Phonology and orthography: The orthographic characterization of rendaku and Lyman’s Law. Glossa: a Journal of General Linguistics 3(1). 1–24.). In Experiment II, I tested whether OCP(diacritic) is psychologically real in the minds of Japanese speakers, using nicknames with [h]→[p] alternation already applied. The results showed that the naturalness of nicknames is reduced when they contain singleton [p] and voiced obstruents, both of which need a diacritical mark in hiragana and katakana. This suggests that OCP(diacritic) is active in nicknaming processes beyond rendaku and devoicing of voiced geminates. Experiment II also showed that the naturalness of nicknames is affected by other OCP effects such as OCP(C), OCP(CV), and OCP(labial). This result suggests that such OCP effects impinge on the patterns resulting from nicknaming formation.
... Formal complexity with respect to the contiguity-similarity tradeoff means that "phonological theory typically treats dependencies between adjacent elements as the normal case, excluding long-distance interactions unless the interacting segments share some property which is absent from intervening material" (Moreton & Pater 2012a: 693) (see also Jensen 1974, McCarthy 1981, Cole & Trigo 1988, Pierrehumbert 1993, Odden 1995, Gafos 1996, Hansson 2001, Frisch et al. 2004, Rose & Walker 2004, Heinz 2010 Formal complexity with respect to contiguity means that a dependency with a shorter distance should be learned better than one with a longer distance 22 . Finley (2011a) found that if participants learned sibilant harmony with a longer distance between the sibilants during the exposure phase, then they would be able to generalize to the one with a shorter distance, but not the other way (see Section 3.3.2). ...
This chapter introduces the literature on artificial grammar that forms the foundation for this study.
... If this is the correct interpretation, the coordination pattern underlying both acoustic outputs[Alternatively, assume that avoidance of overlap truly weakens as the CC profile changes from equal to unequal sonority. This is consistent with work on OCP effects which does indeed suggest that a more refined notion of 'identity' is involved (Padgett 1992, Pierrehumbert 1993, Selkirk 1993, Berkley 1994). In a sequence of segments, identity of their oral gestures would be a necessary but not a sufficient condition for violating the OCP. ...
Linguistic form is expressed in space, as articulators effectconstrictions at various points in the vocal tract, but also in time, as articulators move. A rather widespread assumption in theories of phonology and phonetics is that the temporal dimension of speech is largely irrelevant to the description and explanation of the higher-level or more qualitative aspects of sound patterns. The argument is presented that any theory of phonology must include a notion of temporal coordination of gestures. Linguistic grammars are constructed in part out of this temporal substance. Language-particular sound patterns are in part patterns of temporal coordination among gestures.1
... In essence, then, it is natural to consider Lyman's Law as OCP(diacritic). It may be that OCP is a general cognitive schema to avoid adjacent similar entities (Pierrehumbert 1993;Frisch 2004), which can take both phonological features and orthographic characteristics as its arguments. Alternatively, one could say that ぱ, ば, だ, が (=/p, b, d, ɡ/) form a phonological natural class in Japanese as [+dakuten] sounds, as long as we admit that phonology can have orthographic features like [+dakuten]. ...
This paper argues that phonology and orthography go in tandem with each other to shape our phonological behavior. More concretely, phonological operations are non-trivially affected by orthography, and phonological constraints can refer to them. The specific case study comes from a morphophonological alternation in Japanese, rendaku. Rendaku is a process by which the first consonant of the second member of a compound becomes voiced (e.g., /oo/ + /tako/ ? [oo+dako] ‘big octopus’). Lyman’s Law blocks rendaku when the second member already contains a voiced obstruent (/oo/ + /tokage/ ? *[oo+dokage], [oo+tokage] ‘big lizard’). Lyman’s Law, as a constraint which prohibits a morpheme with two voiced obstruents, is also known to trigger devoicing of geminates in loanwords (e.g. /beddo/ ? [betto] ‘bed’). Rendaku and Lyman’s Law have been extensively studied in the past phonological literature. Inspired by recent work that shows the interplay between orthographic factors and grammatical factors in shaping our phonological behaviors, this paper proposes that rendaku and Lyman’s Law actually operate on Japanese orthography. Rendaku is a process that assigns 'dakuten 'diacritics, and Lyman’s Law prohibits morphemes with two diacritics. The paper shows that a set of properties of rendaku and Lyman’s Law follows from this proposal. However, since some aspects of rendaku and Lyman’s Law are undoubtedly phonological, the ultimate conclusion is that it is most fruitful to recognize a model of phonology in which it has access to orthographic information. Several consequences of the current proposal are discussed.
... Formal complexity with respect to the contiguity-similarity tradeoff means that "phonological theory typically treats dependencies between adjacent elements as the normal case, excluding long-distance interactions unless the interacting segments share some property which is absent from intervening material" (Moreton & Pater 2012a: 693) (see also Jensen 1974, McCarthy 1981, Cole & Trigo 1988, Pierrehumbert 1993, Odden 1995, Gafos 1996, Hansson 2001, Frisch et al. 2004, Rose & Walker 2004, Heinz 2010 Formal complexity with respect to contiguity means that a dependency with a shorter distance should be learned better than one with a longer distance 20 . Finley (2011a) found that if participants learned sibilant harmony with a longer distance between the sibilants during the exposure phase, then they would be able to generalize to the one with a shorter distance, but not the other way (see Section 3.3.2). ...
An important question in linguistics involves the nature of apparent substantive biases. Biases are often claimed to be universal based on typological evidence. This thesis tests for a substantive bias, the proposed universal implicational nasalized segment hierarchy in vowel-consonant nasal harmony, using an Artificial Grammar paradigm. In particular, I address whether a pattern that is predicted by the implicational hierarchy is in fact easier to learn than one that is not predicted or that is indeterminate with regard to predictions. I use a grammaticality judgment wug test paradigm to investigate whether it is easier to make a generalization when a more marked blocker (more sonorant segment) or target (less sonorant segment) is presented during an exposure phase and a less marked blocker (less sonorant segment) or target (more sonorant segment) in the test phase than vice versa. I call this the sonority hierarchy type prediction.
In addition to testing the predictions on the basis of the hierarchy, I also test predictions based on natural classes. The natural class hypothesis predicts that a grammar is more learnable if a new segment (a segment introduced in the test phase but not present in the exposure phase) is of the same natural class as an old segment (a segment introduced in the exposure phase).
The experiment was run with speakers of Min (Taiwan Southern Min), a language with no apparent evidence for sonority classes, using a method based on that of Wilson (2006). Experiments were carried out that allow both the sonority hierarchy type and the natural class hypotheses to be tested, taking individual differences (learner types) into account. The results show that both the sonority hierarchy and natural classes play a role, supporting the claim that it is easier to learn a grammar that exhibits a substantive bias than one that does not.
In conclusion, this thesis suggests that the implicational nasalized segment hierarchy is testable and learnable in artificial grammar learning to some extent and natural classes are psychologically real and actively used by participants in nasal harmony.
... Such behavior is reminiscent of "parasitic harmony" where an assimilation process occurs only if the target and trigger are already similar to one another in some respect. The notion of similarity is developed in work by Pierrehumbert (1993);Frisch (2004);Frisch et al. (2004); among others. In their OT analysis of Korean, Davis and Shin (1999) make use of a constraint, SIMILARITY, to account for the unexpected assimilation of falling sonority /l + n/ clusters. ...
This article investigates consonant gemination in late 19th-and early 20th-century haketía, a now moribund, regional dialect of Judeo-Spanish spoken in northern Morocco since the late 15th century. Some, but not all, consonant clusters arising across a word boundary undergo regressive total assimilation, e.g. [n.n] siudad ninguna 'no city' but [z.n] laz niñas 'the girls'. We present novel descriptive generalizations to show that regressive gemination is sensitive to the degree of sonority distance between the coda and the onset. Evidence of parasitic harmony comes from lateral+consonant clusters, which undergo gemination only if the target and trigger consonants are already similar in some respect. In the framework of Optimality Theory, we formalize syllable contact as a relational hierarchy of *DISTANCE constraints and capture parasitic harmony effects by similarity avoidance, or Obligatory Contour Principle, constraints against adjacent consonants with identical manner and/or place features. These markedness constraints interact with other universal faithfulness and markedness constraints in a language-specific ranking that predicts the attested patterns of regressive gemination. This study lends further support to sonority distance effects and gradient syllable contact in phonological theory and shows that similarity avoidance is also necessary to give a full account of regressive gemination in Moroccan Judeo-Spanish.
... Similarity, according to some authors, is not categorical but gradient, being dependent upon the number of features shared by two or more segments (e.g. Pierrehumbert 1992;Frisch et al. 1997): the greater the number of overlapping features, the stronger the tendency towards dissimilation. 5 The importance of featural contrast for onset cluster well-formedness is witnessed to by the fact that the unmarked complex onset is a stop-liquid sequence, that is a low-sonority non-continuant followed by high-sonority continuant. ...
... We exemplify the restrictions of interest in Table 1 with vowel co-occurrence data from Samoan (Alderete & Bradshaw 2013). Vowel combinations are standardized using Observed/Expected values, which is a common measure used to assess over-(greater than 1.0) and under-representation (less that 1.0) in the lexicon (Pierrehumbert 1993). O/E for all pairs containing the low vowel a approaches 1.0, suggesting that the distribution of these pairs is not restricted. ...
This article contributes to the understanding of gradient phonological patterns by investigating graded vowel co-occurrence in Oceanic languages. In particular, vowel co-occurrence patterns in disyllabic stems are investigated in four languages: Samoan, Tongan, Hawaiian, and Fijian, as well as reconstructed forms in Proto-Oceanic and Proto-Malayo-Polynesian. With some variation in degree, all languages exhibit an over-representation of identical vowel pairs (e.g. i–i), an under-representation of similar vowel pairs (i–e), and no special restrictions on dissimilar vowel pairs (e.g. i–o). These graded restrictions are also subject to order effects in all languages because the dissimilar > similar inequality in frequency is only found in certain orders. Our focus is on documenting the patterns supporting these generalizations so that future theoretical analysis will rest on strong empirical ground. In addition, we propose one such analysis using gradient constraints on parasitic vowel harmony.
... Similarity, according to some authors, is not categorical but gradient, being dependent upon the number of features shared by two or more segments (e.g. Pierrehumbert 1992;Frisch et al. 1997): the greater the number of overlapping features, the stronger the tendency towards dissimilation. 5 The importance of featural contrast for onset cluster well-formedness is witnessed to by the fact that the unmarked complex onset is a stop-liquid sequence, that is a low-sonority non-continuant followed by high-sonority continuant. ...
... 20PLACE (Boll-Avetisyan & Kager, 2014;Kager & Shatzman, 2007). Both studies used21 the Observed/Expected (O/E) ratio(Pierrehumbert, 1993), by which observed counts22 of a consonant pair in a database is compared to the counts expected if consonants23 combined at random. The O/E ratio has been used to illustrate consonant co-2009). ...
Highlights
• If OCP-Labial holds as a gradient constraint, specific labial pairs can be exempt.
• Dutch listeners know of such exceptions, which affect their processing of speech.
• Phonotactic knowledge influenced their segmentation of artificial languages.
• Detailed phonotactic knowledge affects processing when task demands are simple.
• Abstract phonotactic knowledge may affect processing when task demands are complex.
Abstract
Many languages restrict their lexicons by OCP-Place, a phonotactic constraint against co-occurrences of consonants with shared [place] (e.g., McCarthy, 1986). While many previous studies have suggested that listeners have knowledge of OCP-Place and use this for speech processing, it is less clear whether they make reference to an abstract representation of this constraint. In Dutch, OCP-Place gradiently restricts non-adjacent consonant co-occurrences in the lexicon. Focusing on labial-vowel-labial co-occurrences, we found that there are, however, exceptions from the general effect of OCP-Labial: (A) co-occurrences of identical labials are systematically less restricted than co-occurrences of homorganic labials, and (B) some specific pairs (e.g., /pVp/, /bVv/) occur more often than expected. Setting out to study whether exceptions such as (A) and (B) had an effect on processing, the current study presents an artificial language learning experiment and a reanalysis of Boll-Avetisyan and Kager's (2014) speech segmentation data. Results indicate that Dutch listeners can use both knowledge of phonotactic detail and an abstract constraint OCP-Labial as a cue for speech segmentation. We suggest that whether detailed or abstract representations are drawn on depends on the complexity of processing demands.
... In laboratory phonology, Pierrehumbert (1993) experimented with a simple feature-overlap definition of similarity to which Broe (1996) added an information-theoretic refinement discounting redundant features. Frisch (1996) recast these definitions in terms of natural classes, rather than features, and Frisch et al. (2004) demonstrate that the Arabic syllable is best described as involving a gradient constraint against similar consonants in initial and final position, the so-called 'Obligatory Contour Principle' . ...
This dissertation focuses on dialect variation. In the field of dialectology, researchers study the influence of various factors on variation in dialectal speech. Initially, dialectologists were interested in identifying geographical dialect regions. Later, however, they became more interested in the influence of social characteristics (such as age and status). In general, (social) dialectologists have restricted their studies to a small set of linguistic variables (e.g., word pronunciations) which they selected themselves.
Dialectometry, a subfield of dialectology, is characterized by a less subjective approach, in which a large set of linguistic variables is analyzed simultaneously. By counting how many linguistic variables differ between one speaker and another, a measure of linguistic distance between the two speakers is obtained. This measure can be made more precise by taking into account how much individual pronunciations differ (e.g., the pronunciation of ‘can’ is closer to ‘cat’ than to ‘bat’).
In contrast to dialectology, dialectometric studies have lacked a focus on the social dimension and have almost exclusively investigated the connection between dialect variation and geography. By calculating a single distance between two speakers on the basis of a large set of linguistic variables, researchers in dialectometry have also been criticized for their lack of attention to the contribution of individual linguistic variables. Consequently, in this thesis we propose new dialectometric methods integrating social factors as well as allowing a focus on individual linguistic variables.
Given the importance of individual linguistic variables (in our case word pronunciations), our distance measure needs to be as precise as possible. Besides considering the number of different sounds in two pronunciations, it also makes sense to look at which sounds are involved. For example when only a single sound differs between two pronunciations, it should matter if the corresponding sounds are relatively similar (e.g., [o] versus [u]) or very different (e.g., [o] versus [i]). In Chapter 2 we introduce a novel method to automatically obtain sensitive sound distances. Chapter 3 shows that these sound distances make sense acoustically (e.g., the obtained distance between [o] and [u] is lower than the obtained distance between [o] and [i]), and this suggests that our pronunciation distances will also be more precise. In addition, by using these sensitive sound distances we improve our ability to identify which sounds correspond in different pronunciations (such as [w] and [v] in two pronunciations of ‘vinegar’, [wInIg@] and [vInIg@]). This is demonstrated in Chapter 2.
In Chapter 4 we propose a novel method to identify groups of linguistically similar dialects while simultaneously identifying their underlying linguistic basis (in terms of sound correspondences). Besides applying the method to a Dutch dialect dataset in Chapter 4, we also apply it to an English dialect dataset in Chapter 5. In both cases, we find sensible geographical clusters together with their most characteristic sound correspondences.
In Chapters 6, 7 and 8 we propose an integrative approach to simultaneously investigate the effect of geography, several word-related factors (such as word frequency), and various social factors (such as speaker age) on dialect variation at the word level. The wide applicability of this approach (combining mixedeffects regression and generalized additive modeling) is illustrated by applying the method to three different dialect datasets.
Chapter 6 investigates a Dutch dialect dataset. In addition to the importance of geography, we find clear support for the importance of several demographic factors. Communities with a small number of inhabitants or a high average age have dialectal pronunciations more distant from standard Dutch than those with a large number of inhabitants or a low average age. In addition, we observe that nouns (as opposed to verbs and adjectives) and more frequent words are more resistant to standardization.
In Chapter 7 we investigate a Catalan dialect dataset containing dialectal pronunciations of speakers from Catalonia, Andorra and Aragon. As Catalan is not recognized as an official language in Aragon (in contrast to Catalonia and Andorra), this dataset allows us to study the effect of the standard language on dialectal pronunciations. The results clearly show that Catalan dialects are converging towards the standard language (i.e. younger speakers use pronunciations more similar to standard Catalan than older speakers), but only in regions where Catalan is recognized as an official language and taught in school. Consequently, the presence of an official standard language influences dialectal variation.
While Chapters 6 and 7 focus on pronunciation distances, our method also allows us to study lexical differences (e.g., using ‘car’ as opposed to ‘automobile’). Chapter 8 investigates lexical differences between Tuscan dialects and standard Italian. In line with Chapter 6, we observe that more frequent words (in the heartland of Tuscany) are more resistant to standardization. In addition, younger speakers and speakers in larger communities are more likely to use standard Italian lexical forms.
In conclusion, the novel dialectometric methods proposed in this dissertation should be more appealing to dialectologists, as they incorporate social factors and allow a focus on individual linguistic variables. Furthermore, examining a large set of linguistic variables allows us to obtain a more comprehensive view of dialectal variation than by using only a small set of linguistic variables.
... Several different kinds of "phonological similarity" are referred to in the literature, however, and these various types of similarity have diverse consequences for grammar and learning (for a recent overview, see Gallagher and Graff 2012). For example, some studies examine the effects of phonological similarity between lexical items -operationalized as "neighborhood density" -on speech perception and production (e.g., Luce and Pisoni 1998;Vitevitch 2002;Munson and Solomon 2004;Gahl et al. 2012), while other studies consider the similarity between the various potential forms of a lexical item in explaining distributional regularities such as phonotactic restrictions and environments for alternation and neutralization (e.g., Pierrehumbert 1993;Flemming 2004;Frisch et al. 2004;Steriade 2009;Gallagher 2012). Phonological similarity between individual sounds or natural classes of sounds has been measured perceptually via perceptual confusions or explicit mappings with goodness-of-fit ratings (e.g., Miller and Nicely 1955;Strange 1999;Best et al. 2003;Chang 2009b). ...
Although the general notion of “phonological similarity” has figured prominently in linguistic scholarship, the manner in which talkers determine similarity between phonological units is not well understood. Recent research has shown that perceptual similarity does not account for input-output mappings between languages as well as it does within a language. I argue that this disparity arises due to the fact that, unlike phonological similarity within a language, phonological similarity between languages is highly influenced by an abstract, between-system level of analysis that is only relevant cross-linguistically. I review an array of findings in cross-linguistic research to demonstrate, first, a dissociation between acoustic distance and phonemic correspondence and, second, a consistent preference for relating segments and natural classes cross-linguistically on the basis of phonemic correspondence. This type of abstract comparison helps to explain why effects of perceptual similarity are often masked in cross-linguistic circumstances, and I point to possible bases of these comparisons: (1) inventory niches based on contrastive feature oppositions and/or relative phonetics, and (2) distributional parallelisms.
... For example, in Russian the labial consonants form identity classes for the purposes of root cooccurrence restrictions, while the coronals do not; they are split up into stops, fricatives, and sonorants (Padgett 1991(Padgett [1995). Furthermore, examination of the cases presented in Yip 1989, Mester 1986, Pierrehumbert 1993, and Kawasaki 1989 reveals a strong bias towards weaker cooccurrence restrictions on coronals than noncoronals. ...
IntroductionDisharmony outside of Autosegmental PhonologyCoronal Unmarkedness in DissimilationConclusion
... However, their high degree of similarity causes speakers to confuse them, and the more speakers confuse them, the more likely they are to undergo interparadigm leveling. The notion of similarity has been addressed in various contexts, in particular with reference to segmental similarity (Pierrehumbert 1993;Frisch 1996;Cohen 2009). Here we study paradigm similarity, asking what makes two paradigms similar, or more precisely, how do we quantify similarity between paradigms? ...
Paradigm leveling, whereby the number of surface allomorphs within a paradigm is reduced, is often accompanied by the loss of contrast between paradigms (analogy). Since these two are independent of each other, we distinguish between intra- and inter-paradigm leveling. In this paper, we study inter-paradigm leveling in the verb system of Hebrew, manifested by on-going change-oriented variation. In this context, we respond to two questions often addressed in studies on paradigm leveling: (i) Why do some paradigms interact in inter-paradigm leveling and others do not? (ii) What determines the direction of leveling? With regard to the first question, we argue that inter-paradigm leveling is triggered by similarity between whole paradigms, and propose a model that quantifies similarity and predicts the relative chance for two paradigms to undergo inter-paradigm leveling. With regard to the second question, we identify two types of directionality, uni- and bidirectional leveling, and show that the selection between these two is determined by the size of the inflectional classes. Class size determines the direction in unidirectional leveling (the larger is the winner) and class size ratio distinguishes between uni- and bidirectional leveling (the higher the ratio the greater the chance for unidirectional leveling).
... As a measure, we used the Observed/Expected (O/E) ratio (Pierrehumbert, 1993). ...
OCP-PLACE, a cross-linguistically well-attested constraint against pairs of consonants with shared [place], is psychologically real. Studies have shown that the processing of words violating OCP-PLACE is inhibited. Functionalists assume that OCP arises as a consequence of low-level perception: a consonant following another with the same [place] cannot be faithfully perceived as an independent unit. If functionalist theories were correct, then lexical access would be inhibited if two homorganic consonants conjoin at word boundaries—a problem that can only be solved with lexical feedback.
Here, we experimentally challenge the functional account by showing that OCP-PLACE can be used as a speech segmentation cue during pre-lexical processing without lexical feedback, and that the use relates to distributions in the input.
In Experiment 1, native listeners of Dutch located word boundaries between two labials when segmenting an artificial language. This indicates a use of OCP-LABIAL as a segmentation cue, implying a full perception of both labials. Experiment 2 shows that segmentation performance cannot solely be explained by well-formedness intuitions. Experiment 3 shows that knowledge of OCP-PLACE depends on language-specific input: in Dutch, co-occurrences of labials are under-represented, but co-occurrences of coronals are not. Accordingly, Dutch listeners fail to use OCP-CORONAL for segmentation.
... Similarity arises repeatedly as a factor of phonological relevance. Recall, for example, the OCP-Place restrictions in Arabic verbal roots referred to in the previous section (Pierrehumbert 1993;Frisch and Zawaydeh 2001). Within OCP-Place research, place of articulation groupings are further divided; coronals in Arabic, for example, can be broken into sonorants, obstruent stops, and obstruent fricatives, and verbal root restrictions in Arabic most strongly prohibit consonantconsonant sequences that share both place and manner of articulation. ...
While many phonological processes are local, consonant harmony is of interest phonologically because it can occur non-locally. Sibilant harmony in Navajo requires that sibilants within a word have matching anteriority specifications. The process is described as being sometimes mandatory and sometimes optional, but neither the statistical nature of the occurrence in optional settings nor the factors contributing to the optionality are fully understood. This paper provides preliminary investigation into these issues using the first person possessive morpheme, which is underlyingly /ʃi-/ but may harmonize to [si-]. Experiment 1, an online grammaticality judgment survey, reveals that the harmonized prefix is dispreferred in all environments. Experiment 2 presents acoustic data from three Navajo speakers: though none harmonize overtly, the spectral mean and lower bound of frication energy of the prefixal fricative are affected by the presence of [+anterior] sibilants in noun stems. The overall implication of these findings is that harmony is not only optional but is dispreferred or wholly absent for some speakers. While multiple factors are investigated, the only one that consistently affects harmony is adjacency of the trigger and target, indicating that, although consonant harmony may indeed be a non-local process, its occurrence is heavily mediated by distance.
Probability and frequency are becoming increasingly important in phonological analysis. This article reviews contemporary perspectives on how phonological theory addresses gradient phonological patterns shaped by probability and frequency, drawing on theories of the lexicon, grammar, and statistics. After examining their motivations, we show how these diverse theoretical perspectives have been applied to a variety of problems in core phonology, including phonotactics, morphophonology, sound change, phonological categorization, and language development. Our review of theory and applications supports a growing consensus in the field that phonological theories must reckon with probability. Our review also identifies problems stemming from a lack of cohesion in the field, and suggests potential solutions to these problems.
The current study set out to investigate the relative contributions of Initials and Finals to Chinese L2 comprehensibility based on Brown's functional load (FL) principle. 75 speech samples elicited from 20 Urdu-speaking learners of Chinese were subjectively rated by native speakers of Chinese for comprehensibility scores, and then the segmental errors were analyzed based on FL principle. The experimental results showed that the ratio of segment errors with high FL has a stronger correlation with comprehensibility than those with low FL, and the ratio of Final errors showed a stronger correlation with comprehensibility than that of Initial errors, suggesting that segmental errors with high FL inhibit comprehension more than low FL errors, and Finals are more important for successful comprehension than Initials. This study offers: (1) the adaptation of Brown's FL principle on Chinese, (2) an empirical evidence that Final is a more important constituent than Initial in speech comprehension, (3) re-examination of the stronger impact of high FL errors than low FL errors on comprehensibility.
An earlier study [1] illuminated the role of F0 contours on the focus accent in differentiating question and sentence mode in the Sofia variety of Bulgarian.The results showed that placement of the low target of the pitch accent at the beginning and the peak at the end of the accented syllable or in the following syllable (L*+H) is critical for the perception of (syntactically and lexically unmarked) checks, which are used to confirm already known information [2]. By shifting the peak leftwards towards the beginning of the accented syllable (H*) the pragmatic category changed from check to statement. Both categories have a L% boundary tone. However, subjects' judgements indicated that the strength of the intonational information was not equal for the two utterance types. The situational pre context had a strong influence on the interpretation of the utterance as a check or statement. In case the intonation contour diverged from the unmarked contour for a particular function, the category judgement was accompanied by a change in the emotional message.
To examine the general validity of these observations a further experiment was carried out. Three checks, three statements with low terminal boundary tones (L%), and three statements with continuation rises (H%) were selected from Map Task recordings made for a number of male and female speakers [3]. From each of these 9 natural utterances three intonational variants were generated, one for each pragmatic category. Firstly, a stylised resynthesized version of the original (e.g. a check) was produced. Then the intonation contours for the other two pragmatic categories (e.g. statement with terminal fall and continuation rise) were derived from the stylised contour.
Four repetitions of the stimuli were presented (Roman square design) to 15 native speakers of Sofia Bulgarian in three situational contexts: question, neutral statement and polite statement. The context utterance together with the stimulus form a minimal dialogue. The natural context for the check was a statement, and for the two statement forms it was a question. In the test, each context was offered with each pragmatic category, producing potential tension between context and stimulus. The subjects were required to judge - on a five-point scale - the degree to which each stimulus was suited to its context.
The results show that all three intonational contours can be accepted as statements in the context of a preceding question, whereas the change of context cannot shift the interpretation of a statement to a check. The following explanation can be offered for this asymmetry of reinterpretation. The context plays an extremely important role for the interpretation of checks or statements. The context priming a statement (question-answer sequence) provides enough information to uniquely specify the communicative frame. It is a strong enough speech act marker to relegate function of the intonational form to a minor one. Thus the context weakens the distinctive function that intonation has when word sequence and syntactic structure are identical.This does not, however, mean that the intonational form is irrelevant. The shift in the interpretation of the sentence mode (check to statement) can only occur because a compensatory change of modal meaning accompanies it. The check contour cannot be accepted as a neutral statement, it can only be accepted as an emphatic, impatient or angry statement. The statements with a continuation rise were also accepted in the statement context, but the compensatory modal message was of an exaggeratedly polite speaker. Apparently, when a typical (neutral) feature of a particular communicative situation is replaced by a feature typical of another situation, it introduces an additional modal marking. Thus intonation alone, without syntactic and lexical support, can imply a certain shade of modal meaning. This phenomenon is already known at the grammatical level. In Bulgarian, for example, the future tense is the neutral form for referring to events in the future. If the present tense is used, the utterance is immediately modally marked as a firm intention. In English and German the reverse is true.
References
[1] Andreeva, B. & Barry, W.J. (1997). "Intonation von checks in der Sofia-Varietät des Bulgarischen," presented at FDSL2 Potsdam (to appear in Proceedings of FDSL2).
[2] Kowtko, J., Isard, S. & Doherty-Sneddon, G. (1992). "Conversational games within dialogue," Research paper HCRC/RP-31, 1-12. University of Edinburgh.
[3] Anderson, A.H., Bader, M., Bard, E.G., Boyle, E., Doherty, G., Garrod, S., Isard, S., Kowtko, J., McAllister, J. Miller, J., Sottillo, C., Thompson, H. & Weinert, R. (1991). "The HCRC Map Task Corpus," Lang. and Speech 34(4), 351-366.
Theories of the phonology-morphology interface can be differentiated by their claims regarding the timing of phonologically conditioned suppletive allomorphy (PCSA) and phonology. Some (e.g. Paster 2006; Embick 2010) argue that PCSA occurs in a morphological component of the grammar that precedes phonology; others (e.g. Kager 1996; Mascaró 2007; Smith 2015) argue that at least phonologically optimizing PCSA occurs in the phonological component of the grammar, in parallel with phonology. This paper discusses a case of apparently optimizing PCSA in Yindjibarndi (Pama-Nyungan, Wordick 1982), proposes an analysis in which suppletive allomorphy precedes phonology, and shows that the alternative – an analysis in which PCSA occurs in the phonological component of the grammar – should be dispreferred.
This chapter provides a review of evidence that occurrence frequency has an influence on phonological patterns. The examination of quantitative or statistical patterns in phonology, and grammar more generally, pre‐dates the development of modern linguistic theory. For example, Zipf (1965) examined statistical properties of texts and noted a variety of effects that still have relevance in current theoretical discussions, such as the tendency for reduction of high‐frequency words. Bolinger (1961) and Herdan (1962) discussed gradience and statistical distribution as a natural middle ground between the grammatical extremes of the acceptable and the unacceptable. However, given the limitations both in the understanding of language structure and in the computational resources for conducting quantitative studies, this type of research did not gain much traction in the field of linguistics more generally. With the availability of lexical and usage corpora now available, and armed with the descriptive and theoretical advances of modern linguistic theory, a variety of authors are now arguing for the influence of frequency on phonological patterns in synchronic phonology and morphophonology, phonological acquisition, and diachronic phonology. In many ways, phonology is the ideal domain in which to study the potential role of frequency effects in grammar (Herdan 1962). The set of basic phonological units (phonemes or features) is relatively limited, these units are routinely combined in a reasonably static set of fixed forms (the lexicon), and productive combinations of these units in morphology are also limited in their variety of combination by that same fixed set of units. Morphophonological changes are triggered by phonological environments defined by features ( chapter 17: distinctive features ) or phonemes ( chapter 11: the phoneme ) and the results of the changes are within the same set of featural or phonemic varieties.
Dissimilation prototypically refers to a situation in which a segment becomes less similar to a nearby segment with respect to a given feature. As a synchronic alternation, it can be exemplified by liquid dissimilation in Georgian, where the ethnonym‐forming suffix {‐uri} becomes [uli] when an /r/ precedes it anywhere within the word (Fallon 1993; Odden 1994). The resulting pattern of alternation is shown in (1).
The thesis is a psycholinguistic study of Sequential Voicing (known as Rendaku) in modern Japanese. Rendaku is a morphophonemic alternation whereby an initial voiceless obstruent of the second constituent of a compound undergoes voicing. It is infamous for its irregularity, which is attributed to more than a dozen conditioning factors. The study questions its status as a 'productive phonological rule' and investigates the way in which it is acquired and developed by 131 adult and non-adult native speakers. The psychological reality of Rendaku is tested orally by means of an elicited production test and by an acceptability test, both involving nonsense words as compound constituents. Two theories join forces to interpret the results. Optimality Theory provides the most up-to-date analysis of Rendaku, notably of the OCP[+voice, -son] effect (the disfavouring of adjacent voiced obstruents) that blocks the phenomenon. It also enables the construction of a model of multi-staged developmental grammars, each identified by a unique constraint ranking and its outputs. The dual-mechanism model of morphological learning, supplemented by findings on morphological representations and the acquisition of compounds, explains how regular and irregular aspects of Rendaku can be handled differently in the cognitive system. The results show: 1) late acquisition, presumably triggered by the growth in vocabulary and general cognition, 2) low productivity, indicating the weakness of the process, 3) OCP[+voice, -son] as a near-categorical constraint, 4) the distance effect of the OCP, 5) the absence of productive Rendaku for the majority of subjects, 6) psychologically real effects of some phonological conditions, 7) notable individual variations, for instance the preference of voicing certain obstruents. It is concluded that for most subjects Rendaku is not a productive rule of grammar as often described, but is largely a part of generalizations in the lexicon. Different speakers seem to construct different I-linguistic generalizations about Rendaku, which suggests the idea of "parametric poverty" - that is that certain parameters may be set randomly or left unset.
In previous surveys of long-distance consonant harmony, the major place features [labial], [dorsal] and [coronal] are conspicuously absent from the set of possible harmonising features. Ngbaka Minagende displays major place harmony between labial-dorsal segments and simple labials and velars, thus filling this empirical gap. The presence of complex segments with multiple place is crucial to seeing this harmony pattern clearly. These patterns are best handled in the Agreement by Correspondence framework with an active CC-I dent [place] constraint. Other analyses either cannot capture the pattern at all or require fundamental changes elsewhere in phonological theory. The data are supported by a new digitisation and statistical analysis of a Ngbaka Minagende dictionary.
This paper contributes to the typology of laryngeal harmony by analysing an unusual case of long-distance laryngeal co-occurrence restrictions and alternations in Lezgian. This pattern, previously unmentioned in the phonological literature, is the first known case of alternations involving ejective harmony. In Lezgian, local processes mask the interaction of ejectives and plain voiceless stops. This is robustly supported by our dictionary analysis, which reveals a ban on the co-occurrence of ejectives and plain voiceless stops within the foot. Both harmony alternations and static co-occurrence restrictions are sensitive to foot structure, unlike previous cases of consonant harmony. Harmony also interacts opaquely with vowel syncope, and certain co-occurrences of plain and ejective stops are resolved with dissimilation rather than harmony, showing a conspiracy to avoid co-occurrences. We demonstrate an account within the Agreement by Correspondence framework and discuss implications for the typology and analysis of consonant harmony.
Morpheme structure constraints (MSCs or, equivalently, morphemic constraints), i.e., constraints on underived items, pose a logical problem to any theory claiming that the first input list to the lexicon, i.e., the dictionary (DICT), may contain material which is considered ill-formed by the MSCs of the language. This problem is related to Scobbie’s (1991:1) “interaction problem”, i.e., the need to make explicit the way in which constraints, rules and representations interact. In the case of MSCs, the interaction problem can be stated as follows: how can ill-formed material be present in the DICT if the DICT contains constraints, i.e., MSCs, against ill-formed material? The Theory of Constraints and Repair Strategies (TCRS) proposed by Paradis (1988a, 1988b, 1990, 1993, this volume), along with Optimality Theory (e.g., Itô and Mester, this volume, and McCarthy, this volume) but in contrast with Declarative Phonology (e.g., Scobbie, this volume), claims explicitly that the DICT can contain material which is identified as ill-formed by post-morphemic constraints (see Paradis and Prunet 1989a:331; Bagemihl 1991:641; Paradis and El Fenne 1991, 1992, 1993; and Ulrich 1991 for other instances of underlying ill-formedness). To solve this interaction problem, Paradis (1993, this volume) suggests that phonological constraints are active in the lexicon (the component where words are morphologically derived) and the post-lexical level alone, not in the DICT. Structures of the DICT may be identified as ill-formed only when going through the lexicon. The theory of phonology which emerges from this position clearly contradicts the commonly-held view that DICT entries are constrained by MSCs.
This article investigates the role that the phonetic parameters of duration and voicing play in shaping asymmetric patterns of Romance stop-liquid cluster realization. Based on acoustic analysis of experimental data from Quebec French and Argentine Spanish, we demonstrate the existence of an asymmetry in the proportional duration of the stop and a following epenthetic vowel or lateral: sonorants are shorter after voiceless stops in stop-liquid clusters. Rhotics do not participate in this process. The Spanish tap does not vary significantly in length, and the French dorsal fricative is longer in voiceless clusters. We propose that compensatory adjustments of the following sonorant are the result of the interaction of coarticulatory constraints, both universal (shorter sonorants after voiceless stops or fricatives) and language-specific (voicing in dorsal fricatives in French). Parallels are drawn between the synchronic variation attested and well-known patterns of diachronic change in Romance.
This paper reports on a pattern of voicelessness dissimilation in the Kordofanian language Moro. Voiceless stops and affricates become voiced before a voiceless obstruent in a transvocalic configuration. The dissimilation is robust, and productive across morphological contexts. Phonetically, voicing in Moro is realised as a difference between prevoicing and short-lag voice onset time. This makes [voice] the most realistic featural characterisation; using another feature like [spread glottis] in lieu of [–voice] doesn't explain the contrast. Consequently, dissimilation of voicelessness in Moro is strong evidence that [voice] is a binary feature, and that […voice] may be phonologically active despite being ‘unmarked’. We show that when [–voice] is admitted, the Moro pattern is straightforwardly analysed on a par with other cases of dissimilation. Our analysis uses the theory of surface correspondence, which carries no crucial assumptions about markedness; other theories of dissimilation are considered in an online supplement.
Wm. G. Bennett (2015). The phonology of consonants: harmony, dissimilation, and correspondence. (Cambridge Studies in Linguistics 147.) Cambridge: Cambridge University Press. Pp. xix+394. - Volume 33 Issue 3 - Juliet Stanton
The internal structure of the syllable has been a matter of long-standing debate. Some theories propose a highly articulated tree structure whose topmost constituents are the onset and the rhyme. However, much recent work in phonological theory has adopted a flatter model in which neither the onset nor the rhyme constitutes a constituent.
Experiments using the novel word game paradigm developed in Treiman (1983) have been interpreted as supporting the onset-rhyme model. Here we present new results using this paradigm. Subjects learned to insert infixes into simple monosyllabic words, and then extended the infixation to more complex forms including longer words with variable stress placement. The results are interpreted in the light of findings about morphophonemic processes occurring in natural language. Our model draws on the concept of template mapping, adapted from the literature on prosodic morphology. The patterns observed in the data are better modeled by mapping onto output templates than by any derivational rule referring to onset-rhyme constituency. Though the output templates do include prosodic detail, the level of detail available in flat models of the syllable is sufficient to explain the results. A critical appraisal of these results in relation to other results in the literature leads us to reject the onsetrhyme model.
Similar Place Avoidance (SPA) is a phenomenon well-known to adult languages, where consonants within a root that share the same place of articulation are avoided. The existence of SPA across languages is so robust that it has been claimed to be a statistical phonological universal. Given these claims to universality, this study aims to investigate whether there are any “homorganicity” effects present in the speech of children; i.e. effects that are sensitive to place of articulation. While children often exhibit a stage of consonant harmony, there has been virtually no research involving possible gradient patterns of SPA for children (where the pattern is typically gradient for adult languages). The findings are telling: there is evidence for a homorganicity effect, but one that is driven by agreement in place of articulation, and not avoidance of similar place. This finding is linked to later stages of consonant harmony in child speech.
This volume is a welcome addition to the literature on Italian dialectology, specifically phonology, and to the wider discourse of phonological theory. 1 Originally intended as a companion volume to SyntactictheoryandthedialectsofItaly (Belletti 1993), this volume grew out of a project initiated by the editor in 1995. The contributions collected in this book display the enormous richness of Italy's complex linguistic situation, a complexity which arises from the interaction between Standard Italian and Italian dialects, as well as from influences from other minority languages spoken throughout the Italian peninsula. Standard Italian as a widely used national language is little over 100 years old. However, as Berruto (1993: 3) points out, (my translation). Italian dialects are classified geographically into three macro-categories: Northern (those above the La Spezia-Rimini isogloss), Tuscan and those of the Centre-South (which includes Sicily and Sardinia) (see Lepschy & Lepschy 1992, Sobrero 1993 and Maiden & Parry 1997 for detailed information pertaining to both Standard Italian and Italian dialects). As well as providing descriptive and theoretical accounts of aspects of Standard Italian phonology, the contributions in this volume cover a range of southern and northern dialects, including Friulian, a Tuscan variety (Pisan) and data from the Raeto-Romance of Eastern Switzerland. There are, however, no in-depth treatments of the dialects of Sicily or Sardinia. Theoretical frameworks include classical generative phonology, (early) Optimality Theory, derivational constraint and repair models, moraic phonology and feature-based models. The collection resembles a who's who of Italian phonology, in that it gathers together work from many important established and upcoming scholars in the field.
ResearchGate has not been able to resolve any references for this publication.