Tonal Melodies Research Articles

On Tone and Morphophonology of the Akan Reduplication Construction

Reduplication in Akan has received some discussion in the literature but all the studies have concentrated on some aspects of segmental processes that operate on the base to generate the output. In this paper, we study the morphological, segmental and tonal processes related to reduplicative construction in Akan. We demonstrate that on the basis of tonal perturbations which bases and reduplicative templates undergo, the output tone melody of the reduplicated form vis-a-vis the tone melody of base, we are able to tell the base from reduplicant in the Akan reduplicative structure. We argue in the central portions of this paper that the reduplicant in Akan could be either prefixed or suffixed to the base and, in course of further reduplication construction, it could be sited within the two constituent tokens of the original reduplicative output which serves as an unmarked base for further reduplication. This piece of information counterexemplifies the assertion in the existing literature that in the Akan reduplication construction, the reduplicant is invariably prefixed to the base. In this paper, we study reduplication of verbs, adjectives, nouns, and lexical reduplication demonstrate that words belonging to the same class behave tonally same.

Probabilistic Models of Melodic Interval

Two probabilistic models of melodic interval are compared. In the Markov model, the “interval probability” of a note is defined by the corpus frequency of its melodic interval (the interval to the previous note), conditioned on the previous one or two intervals; in the Gaussian model, the interval probability is a simple mathematical function of the size of the note’s melodic interval and its position in relation to the range of the melody. In both models, this interval probability is then multiplied by the probability of the note’s scale degree to yield its actual probability. The two models were tested on four corpora of tonal melodies using cross-entropy. The Markov model yielded a somewhat lower (better) cross-entropy than the Gaussian model, but is also much more complex, requiring far more parameters. The models were also tested on melodic expectation data, and on their ability to predict the distribution of intervals in a corpus. Possible ways of improving the models are discussed, as well as their broader implications for music cognition.

The Correspondence between Linguistic Tone and Musical Melody

Proceedings of the Ninth Annual Meeting of the Berkeley Linguistics Society (1983), pp. 148-157

Effects of Metrical Encoding on Melody Recognition

We report two experiments exploring whether matched metrical and motivic structure facilitate the recognition of melodic patterns. Eight tonal melodies were composed from binary (four-note) or ternary (three-note) motivic patterns, and were each presented within a metrical context that either matched or mismatched the pattern. On each trial, participants heard patterns twice and performed a same-different task; in half the trials, one pitch in the second presentation was altered. Performance was analyzed using signal detection analyses of sensitivity and response bias. In Experiment 1, expert listeners showed greater sensitivity to pitch change when metrical context matched motivic pattern structure than when they conflicted (an effect of metrical encoding) and showed no response bias. Novice listeners, however, did not show an effect of metrical encoding, exhibiting lower sensitivity and a bias toward responding “same.” In a second experiment using only novices, each trial contained five presentations of the standard followed by one presentation of the comparison. Sensitivity to changes improved relative to Experiment 1: evidence for metrical encoding – in the form of reduced response bias when meter and motive matched – was found. Results support the metrical encoding hypothesis and suggest that the use of metrical encoding may develop with expertise.

Tonal melodies in the Lulamogi verb

In this paper I describe the tonal melodies of Lulamogi, a small understudied language of the Luganda-Lusoga cluster. Although Lulamogi has lost lexical contrast on verbs, as in most Bantu languages, it has a rich set of verbal melodies that are assigned on the basis of tense, aspect, mood, polarity and clause type (e. g. main vs. relative). Although focus is on the stem tone melodies, other complications arise from preceding prefixes which may shift their H tone onto the stem initial syllable. Since the system is quite different from that found in well‑known Luganda, it is of particular interest for our understanding of the evolution of the tone systems of this subregion.

Melodic tone in Totela TAM

Most Totela tense, aspect, and mood forms take one of three patterns of melodic H tones. In the first pattern, only lexical Hs are present. In the second, a melodic H surfaces on the first stem mora, barring interference from general tone constraints such as Meeussen’s Rule. In the third pattern, the melodic H surfaces on the penultimate mora of the stem. Melodic H assignment is also conditioned by prohibitions against two adjacent Hs in the output. All tone melodies in Totela are affected by across‑the‑board application of H‑tone anticipation, in which Hs shift one syllable to the left. In this paper, the three major tone patterns are presented, along with their variations and interactions with other tonal processes. Several minor patterns are also described.

Predictable tone in Bena

Verbal tone in Bena is characterized by a single High tone per word. Roots are toneless, and inflected forms are assigned one of four tone melodies : Pre‑Stem Initial (PSI), Macrostem Initial (MSI), Antepenultimate (APU), and Penultimate (PU). The location of H in an inflected form is entirely predictable through a combination of morphophonological criteria, including the morphological complexity of the verbal base, the tense and aspect of the verb, the form of the final suffix, and whether the verb is affirmative or negative.

Tonal Processes in the Setswana Verb

AbstractThe present article offers an account of how tonal contours of individual verbal forms are generated in Setswana. Its main proposal is that verbal forms have complex structure with internal boundaries that are relevant to tonal processes. The framework of the account, Distributed Morphology (DM), distinguishes between word-building operations that take place in Syntax, in Morphology, and in Phonology, and it is argued that constituency created in the course of these operations is reflected in the tonal melody of individual forms. The present analysis accounts for tonal contours of all forms that posed difficulty under previous treatments while presenting a number of significant theoretical advantages. Most importantly, it offers an insight into Setswana clause structure and defines in precise terms the role that tone plays in making morphosyntactic and prosodic structures phonologically explicit.

Cilungu tone melodies: a descriptive and comparative study

Cet article fournit une description détaillée des schèmes tonals mélodiques hauts (MH) du cilungu. Le schème MH apparaît de quatre façons différentes dans les formes verbales conjuguées ainsi que dans une variété de noms. Après la présentation des données et schèmes cilungu, je souligne les similarités et les différences que ceux-ci présentent avec cinq langues proches appartenant à divers sous-groupes de la zone M.

Bakweri tone melodies

This paper reports on verb tone melodies in Bakweri, which has 6 inflectionally‑governed stem tone patterns. Depending on tense, a verb may have no tones beyond those in the lexicon, or may add L, H, HL, or LH to the stem. These melodic tones are linked to the final vowel, and melody‑initial H spreads leftward to the lexical tone of the root. In addition, the reflexive contributes a LH melody which partially replaces the inherent melody determined by a verb’s tense, and which undergoes a special leftward spreading rule.

Interacting tonal processes in Susu

Susu, like a number of other Mande languages, displays tonal melody neutralization in some nominal constructions. This paper illustrates that in addition to the lexical tonal melody of input elements playing a key role in the outcomes of neutralization, prosodic structure and interactions between multiple tonal processes are also important factors driving observed surface forms. We present novel data on a wide variety of compounds containing complex word shapes; tonal outcomes in such compounds offer a deeper look at factors influencing neutralization. We posit that tones in Susu nominal compounds are subject to Neutralization-by-Copy, which subsequently interacts with analogous, more generalized processes that further smooth the overall tonal contours of a given word.

A prosodic perspective on the assignment of tonal melodies to Arabic loanwords in Bambara*

A prosodic perspective on the assignment of tonal melodies to Arabic loanwords in Bambara*

Exploring the influence of pitch proximity on listener’s melodic expectations.

Two studies, one correlational and one meta-analytic, were conducted to explore whether pitch proximity influences listeners' melodic expectations about pitch direction and tonality. Study 1 used a probe-tone task. Specifically, listeners heard fragments of tonal melodies ending on intervals of 1 or 2 semitones, and rated how well individual tones (the probe-tones) continued each fragment. Regression analyses showed that listeners expected probe-tones to be proximate in pitch to the last tone they heard in the fragments, but not to the penultimate one, as there was no evidence of expectations for a change in pitch direction. However, listeners expected probe-tones to be proximate to (at least one of) the other tones they had heard in the fragments, as there was evidence of expectations for a melodic movement toward the bulk of the fragments' pitch distribution. In addition, the most stable probe-tones in the key of the fragments were more expected than the least stable ones only when they were proximate to the tones presented in the fragments. The results of Study 1 were replicated and extended in Study 2, in which a meta-analysis of data reported in Schellenberg (1996, Experiment 1) was performed. These data had been collected using the same probe-tone task as in Study 1, but different melodic fragments; the fragments ended on intervals of 2 or 3 semitones. Together, the present findings suggest, first, that when small intervals occur in a melody, pitch proximity has only a global influence on expectations about pitch direction; and second, that pitch proximity constrains the influence of tonality on melodic expectation.Keywords: pitch proximity, pitch direction, tonality, conceptualization, statistical and heuristic learningPitch Proximity and Melodic ExpectationThe generation of expectations about what events will occur next in a musical piece and when these will appear has long been thought to be a critical aspect of music listening. Basi- cally, the idea is that expectations that prove to be correct facilitate music understanding, or vice versa, and that the in- terplay between expectations' negation and resolution regulates the emotional content of musical experience (Huron, 2006; Meyer, 1956; Narmour, 1990). During the last decades, this idea has stimulated a great deal of theoretical and empirical research, which has been particularly fruitful in the domain of pitch-related melodic expectation. Research in this domain re- vealed that to a large degree listener's expectations can be accounted for by three factors, namely, pitch proximity, pitch direction, and tonality (Bharucha, 1996; Carlsen, 1981; Huron, 2006; Larson, 2004; Lerdahl, 1996, 2001; Margulis, 2005; Meyer, 1956, 1973; Narmour, 1990; Schellenberg, 1997; Tem- perley, 2008). The present research further investigated these factors by exploring how and to what extent pitch proximity influences expectations about pitch direction and tonality.The influence of pitch proximity refers to the effect of distance in pitch between sequentially presented tones on the listener's expectations. It is understood as a gestalt-like audi- tory principle, perhaps the most basic one, regulating what tones listeners expect to occur next in a melody to be grouped with the tones heard previously (Larson, 1997; Margulis, 2005; Meyer, 1956; Narmour, 1990; see also Bregman, 1990). In line with this, there is strong evidence that listeners expect the next tone in a melody to be proximate in pitch to the last tone they heard; all things being equal, the nearer in pitch the ensuing tone is to the last one, the more highly it is expected (Carlsen, 1981; Cuddy & Lunney, 1995; Eerola, 2004; Krumhansl, 1995; Larson, 2004; Schellenberg, 1996). For example, given the melody shown in Figure 1, at the beginning of measure 4 (see the ? mark in the figure), the expectation for Bb4 would be stronger than the expectation for B4, because the former is nearer to A4 (the last tone in the melody) than the latter; the most expected tone would be A4, the unison. …

Tonosyntax and reference restriction in Dogon NPs

Lexical tone melodies of words and word strings in Dogon languages are overridden by syntactically conditioned word-level or broader tone overlays. In Ben Tey and other Dogon languages, certain NP-internal elements—possessors, adjectives, relative clauses, demonstratives—are controllers, imposing overlays such as {L} and {HL} on adjacent words or word strings (the targets). Other elements (nonsingular quantifiers, discourse-functional morphemes) do not control overlays. The semantic generalization is that reference restrictors are controllers, while non-reference restrictors are not. Tonosyntactic processes have access to input stem-class categories and semantic/pragmatic groupings thereof, input phonological features (tone of the adjacent edge of the controllers), prosodic weight of the target, and, with possible minor exceptions, linear adjacency. Such operations can only be accounted for in a model of grammar with an enriched syntax-phonology interface.

The Influence of Tonal and Atonal Contexts on Error Detection Accuracy

Music education students ( N = 21) at a university in the southeastern United States took an error detection test that had been designed for this study to determine the effects of tonal contexts versus atonal contexts on the ability to detect performance errors. The investigator composed 16 melodies, 8 of which were tonal and 8 of which were atonal. The test administered included 18 planned errors, 9 in tonal melodies and 9 in atonal melodies, and errors were balanced between the two sets of melodies for error duration and error interval. These melodies were performed live for the participants, who were asked to identify where errors took place in the performance. Participant test scores were calculated, and mean scores were generated for each individual error. Each error was then classified according to its attributes of tonal context, interval deviation, and duration. Interval deviation and duration did not significantly affect error detection scores, while tonal context did. More specifically, errors in tonal contexts that deviated from the tonal framework were significantly more easily detected than atonal errors or errors in tonal contexts that remained within the tonal framework. These results confirmed findings in existing research.

Optimizing the relation between tone and prominence: Evidence from Franconian, Scandinavian, and Serbo-Croatian tone accent systems

This paper provides a diachronic explanation for related changes in the tonal melodies of Franconian, Scandinavian, and Serbo-Croatian tone accent systems. It is argued that in each of these systems, some dialect groups have optimized the relation between tone and prominence by replacing cross-linguistically marked low level or rising pitch contours in (nuclear) accent syllables with unmarked high level or falling ones. Two diverse adaptation strategies can be identified, essentially leftward shifts of pitch vs. pitch raising at the beginning of accent syllables. These diverse strategies can lead to a reversal of tonal melodies across related dialects, as can most clearly be demonstrated on the basis of a tonal reversal in declaratives between so-called Rule A vs. Rule B dialects in Franconian. The analysis allows for a reconstruction of the original accent oppositions in Franconian and Scandinavian dialects: it can be shown that despite their different origins, the languages had similar original tonal contours for the two tone accents, and the diverse melodies among modern dialect groups are the outcome of similar developments. The paper contributes to our understanding of diachronic (Germanic) accentology and provides a tool set for reconstructing diachronic changes in the tonal melodies of prosodic systems.

Contextual Evidence for the Representation of Pitch Accents in Standard Serbian

This paper reports the results of an experiment that elicits contextual effects on Rising and Falling accents in Standard Serbian, with the goal of determining their acoustic correlates and their phonological representation. Materials systematically vary the distance between pitch accents, inducing "tone crowding," in order to identify the phonetic dimensions that consistently distinguish the two pitch accent types, to examine the association between accents and the segmental string, as well as the timing relationship between accent minima and maxima, and to investigate the interaction between lexical accents and boundary tones. On the basis of the phonetic findings, a unified analysis of the phonological distribution and phonetic realization of Falling and Rising accents in Standard Serbian is proposed. It is proposed that both Rising and Falling accents consist of a single lexical High (H). The restricted distribution of the two accents emerges from the interaction of stress and tone: Falling accents are monosyllabic, such that stress and pitch prominence coincide; Rising accents are bisyllabic, such that the stressed syllable precedes the pitch-accented syllable. The phonetic differences between the Falling and Rising accents follow from the place of lexically designated H, the location of stress, and the effects of boundary tones. The larger issue we address concerns the phonological characterization of tone/stress interactions. Given the two general types of interactions, one in which the place of stress is predictable from the place of tone, and the other with the reversed direction of influence, we analyze Standard Serbian as belonging to the former type. While both types can be characterized in systems of tonal phonology, which allow free interaction of tone and stress, the type exemplified by Standard Serbian, with contrastive tonal specifications governing the distribution of stress, cannot be captured in an Autosegmental-Metrical (AM) framework, in which stress serves as anchor for tonal melodies.

Timbral and melodic characteristics of Persian and Kurdish singing

The floridly ornamented vocal technique in the courtly heritage of the Persian singing style called Avaz was studied along with excerpts from the flamboyant variety of the vivid Kurdish tradition. Audio and EGG signals were recorded from professional male tenor singers singing stylistically typical song excerpts from each tradition. Voice source parameters and formant settings (F1 & F2) were measured from inverse filtering of the audio signal, using the custom made DeCap (Svante Granqvist) and the commercial Soundswell softwares. Fundamental frequency F0 was measured from the EGG signal using the Soundswell CORR tool. In all melismatic embellishments, melody tones were preceded by short falsetto episodes whereby F0 quickly jumped up to a peak. For example, rapid tone repetitions as well as drill-like alternations between two neighbouring scale tones were interleaved with short falsetto segments. Moreover, for most vowels, the singers tuned F1 and sometimes also F2 to a spectrum harmonic in the higher part of their voice range, i.e. above about Bb4 (235 Hz, approximately). These findings will be discussed in relation to other singing styles, such as western operatic singing.

Speech rhythm in Akan/Twi: A preliminary experiment

This study explores the rhythmic timing patterns of Akan/Twi, a West African tone language thought to be syllable-timed, using the Speech Cycling Task (Cummins & Port, 1998; Port, 2003, 2007; Tajima & Port, 2003). How rhythm appears in tone languages without stress/accent is poorly understood. In the Speech Cycling Task, speakers were expected to place prominent elements within rhythmic modes, where specific syllables are pronounced more frequently in certain temporal regions. To uncover which syllables are considered rhythmically prominent, a previous tapping experiment (Darwin & Donovan, 1980; Donovan & Darwin, 1979; Purvis, 2009) was used in which subjects displayed entrainment between tapped beats and specific syllables (akin to stress-timing), rather than beat entrainment with all syllables. Prominent elements should also be resistant to temporal displacement when syllables are inserted between them and temporal compensation should occur when syllables can be deleted. The data included 20 phrases ranging from 4 to 8 syllables in length and four tone melodies (H, L, HL, and LH) repeated at varying rates. Results suggest subjects prefer certain rhythmic modes and that displacement and compensation occur such that rhythmic patterns change. Implications for the stress-timing/syllable-timing dichotomy and how tonal melodies affect rhythmic patterns are discussed.

Special Tonal Alternations in Dongshi Hakka Reduplication

This paper investigates the special tone sandhi phenomena in the reduplication patterns of AB(C)AB(C), AAA, and AAAA in Dongshi Hakka. These patterns are notable in that they involve tone sandhi patterns that are different from those found in non-reduplications. Within the constraint-based theory of OT, this paper shows that although the tonal alternations observed in the different reduplication patterns are distinct from one another, they are all accompanied with a floating high tone which contributes greatly to the special patterns associated with tone sandhi. In addition to the existence of a floating high tone, AAA reduplication and AAAA reduplication also surface with fixed tonal melodies. It is shown that the fixed tonal melodies in the two patterns of reduplication actually resemble the common phenomena of fixed segmentism in reduplication and are driven by markedness requirements.