Latest articles from "Journal of Singing":


Settle the Score: Performer's Analysis for the Classical Singer(May 1, 2015)


All the World's a Stage(May 1, 2015)

Graduation Advice: Doing versus Having(May 1, 2015)

An AIDS Quilt Songbook: Sing for Hope(May 1, 2015)

NATS and the NATS Foundation-Working Together for You(May 1, 2015)

Other interesting articles:

Birth control: Power/knowledge in the politics of birth
Health Sociology Review (December 1, 2014)

Johann Adolph Scheibe
Music Library Association. Notes (March 1, 2015)

The Motet around 1500: On the Relationship of Imitation and Text Treatment?
Music Library Association. Notes (March 1, 2015)

The Cambridge Companion to the Symphony/The Cambridge Wagner Encyclopedia
Music Library Association. Notes (June 1, 2015)

Subject of Passions: Charles Le Brun and the Emotions of Absolutism
Philological Quarterly (January 1, 2014)

Modern Drummer : MD (May 1, 2015)

Ridicule and Wonder: The Beginnings of Minstrelsy and New York
Afro - Americans in New York Life and History (July 1, 2012)

Publication: Journal of Singing
Date published:
Language: English
PMID: 78638
ISSN: 10867732
Journal code: JRLS

(ProQuest: ... denotes non-USASCII text omitted.)

(ProQuest: ... denotes formulae omitted.)

THE MUSICIAN WHO HAS MADE a study of lyric diction will have come across discussions of syllabification, as it applies to various languages. One might well ask whether an awareness of how words syllabify has any practical relevance to singing. Is it merely textbook analysis of a linguistic policy whose realization is obvious from any text underlay in vocal music? Or are there compelling reasons why a performer should be concerned with the manner in which syllables are organized? This article will consider specific features of syllabification in Italian, French, German, and English, with a view to justifying their importance.1

Although everyone has an intuitive understanding of what a syllable is, an accurate and adequate definition of the term has proven to be remarkably elusive for linguists. Our instinct about what constitutes a syllable derives from an awkward amalgam of acoustic, articulative, grammatical, and perceptual observation, and any attempt to define the syllable along the lines of any one of those parameters inevitably falls short of both analytic precision and our own intuitions. For our purposes, we need not be detained by the theoretical niceties of the linguistic arguments that have been promulgated. Our examination will be based on the most widely recognized analytic methodologies, discussing acoustic, articulative, grammatical, and perceptual considerations as appropriate.

Apart from individual letter symbols, words can be dissected into their segmental components in several ways. A word can be considered to be a string of discrete phonemes within a language, like pearls on a strand, or as a complex flow of articulative motions occurring simultaneously. The former is the traditional approach that is axiomatic to IPA transcription methods. The latter comes closer to reality, but lacks a clear nomenclature matrix or graph that does full justice to the complexity of human utterance, while remaining visually simple. Words can also be dissected into their constituent morphemes, or smallest units of meaning. Thus, inextricably splits into four morphemes: the negation prefix in-, the root stem -extric- (from extricate), the adjectival suffix -able, and the adverbial suffix -ly.2 But when it comes to dividing the word into syllables, more than one possible rendering seems plausible. It is not always appreciated or made clear that one can speak of syllabification on two quite distinct levels, orthographic and phonetic. Let us consider orthographic first. At least five potential realizations present themselves:

in-ex-tric-ab-ly (four closed syllables, one open)

in-ex-tric-a-bly (three closed syllables, two open)

in-ex-tri-ca-bly (two closed syllables, three open)

i-nex-tri-ca-bly (one closed syllable, four open)

i-ne-xtri-ca-bly (five open syllables)

Further combinations could be devised from other combinations of the above. Who decides whether a syllable should be closed or open in English? Both are commonplace, and often more than one option is possible. As a rule, consonant clusters at syllable junctures will divide to reflect the phonotactic restrictions that prevail in the language. In other words, if a cluster is not permissible in onset or in coda position, such a cluster will not be grouped together orthographically in syllable division. By this rule, i-ne-xtri-ca-bly would not be on the list. Furthermore, the converse of the rule is not the case. Phonotactic permissibility is a necessary but not sufficient criterion for grouping consonants together in onset or coda position. In our model word for instance, /tr-/ is a common onset cluster, but so is /str-/ in English. The letter <x> of course suffices for two phonemes, /k/ and /s/, and cannot be split for orthographic syllabification. If one considers phonetic syllabification however, ... or ... is likely to result, depending on which syllable is deemed to have the primary stress. No pronouncing dictionary splits the phonetic syllable after the ... in spite of the fact that /-ks/ is a permissible coda (box, locks), and /tr-/ a permissible onset. Intuitively we feel that ... is "more natural," probably because /k/ is velar while ... are all alveolar. As we have seen, ... is not an option because [ks-] is not permissible in onset. Elsewhere in the word, ... is preferred to ... is preferred to ... and ... is preferred to .... This is done in spite of the fact that all options are common phonotactically except for the last (syllables do not end in ... in English). Other factors come into play regarding our intuitive sense of which is "better." For instance, ... reflects the morphological structure of the word better than ... The latter also forces ... to stand alone as a <V> syllable - not a typical occurrence in English. The same would be true of the ... in ... The guidelines that govern English phonetic syllabification are thus involved, but in most cases it will be found that the standard phonetic syllabification will most closely approximate the orthographic syllabification that has the fewest closed syllables.3

Dictionaries adopt different lexicographic conventions in the listing of words and their pronunciations. Headwords are often syllabified in American English dictionaries, but less so in British English dictionaries. Some dictionaries indicate only word divisions that are eligible for hyphenation in typographic word-wrapping. Other dictionaries indicate all boundaries, whether they can serve as linewrap divisions or not (gar-gan-tu-an), thus conflating two separate considerations rather confusingly. Such divisions are typically oriented toward morphological or etymological clarity, rather than phonetic considerations. Just as phonetic and orthographic syllabification necessarily diverge sometimes, hyphenation for the purposes of line wrapping may differ again from both. When orthographic syllabification differs from phonetic, it is through the influence of morphological/etymological considerations. Words sometimes have different pronunciations, stress patterns, and syllable counts, which complicates the lexicographer's task further. Vocal scores print texts according to standard rules for orthographic syllabification, while IPA transcriptions follow phonetic principles in the syllabifying of text. The two approaches do not always coincide precisely, and in French especially, the distinctions are considerable.


Although alternative methods have been suggested, linguists usually analyze syllables according to phonetic considerations. Syllables are traditionally considered to contain up to three parts:

1. the onset (consonant, consonant cluster, or semi-consonant);

2. the nucleus (usually vocalic; some languages, like English, German, and Czech, possess nuclear consonants);

3. the coda (consonant, consonant cluster, or semi-vowel).4

A syllable may or may not contain an onset or coda, but cannot exist without a nucleus. Possible combinations of consonants and vowels in syllables include

V a, ear, eye, owe

CV by, hair, see, why, you

VC at, of, ought, out

CVC but, heart, tooth, thought

CCV clay, spa

VCC ask, oils

CCVC clash, plate

CVCC leaves, task, walks

CCVCC clothes, grasp

and so on, up to CCCVCCCC, at least in English (strengths). In analyses such as this, V is understood to encompass pure syllabic vowels, diphthongs, and triphthongs, while C includes onset semiconsonants like [j], [w], and .... If a syllable ends in a vowel or semivowel it is "open," and it is "closed" if it ends in a consonant. Note that C and V are defined phonetically, not orthographically, in the analysis of syllable. The word banana consists of three syllables with identical structure. The state Iowa consists of three syllables of the form V + V + CV, in which the <w> is "ambisyllabic"; that is, it signals the semivowel that ends the second-syllable diphthong, and functions as an onset semiconsonant [w] to the third syllable. Many consonants at syllable boundaries are ambisyllabic in English, especially in words where morphological and phonetic considerations differ.

Each language has its own set of rules regarding syllable structure. They are closely related to phonotactics-the study of the allowable sequences of consonants and vowels within a syllable, word, and phrase. Languages place different constraints upon permissible sequences of phonemes, and foreign language acquisition requires learning to articulate new sequences, while avoiding familiar ones that are not allowed in the new language. A familiar example is German, which allows onset [pf-], [kn-], [∫p-], and other clusters foreign to English, but forbids English [sp-], [st-], etc. in onset. Non-European languages often have fundamentally different approaches to syllable structure. For instance, all Japanese syllables are open, except for coda [n].5 Chinese adopts a syllable-based (rather than phoneme-based) orthography, in the form of ideograms, which makes it more difficult for native Chinese speakers to comprehend the segmental components (vowels, consonants) of a syllable which alphabetic writing systems reflect. Although it is tempting to think of the alternation of vowels and consonants as a language universal, it is physiologically feasible to utter strings of consonant clusters that would be forbidden in most languages.6 The native Salishan languages of western North America (such as Nuxálk, or Bella Coola) contain bewilderingly long clusters of voiced and unvoiced consonants without intervening vowels, thus calling into question the very notion of a syllablebased analysis of language.

While syllable nuclei are typically vocalic, this is not inevitably so. English allows syllabic nasals and liquids, as in sudden ... It is debatable whether the final <m> of words like schism and chrism constitutes an extra syllable [-z.m], or simply a coda cluster [-zm]. In music, they are treated as clusters within monosyllables. Czech allows syllabic [-r-] and ... as in the unlikely but nationally popular tongue-twister, Strc prst skrz krk ("Stick your finger through your throat").7

Modern approaches to phonology organize the characteristics of speech according to a wide spectrum of features, one of which is "sonority." Acoustic phonetics ranks phones according to their level of sonority, or loudness. An index comparing the relative "carrying power" of groups of phones has been proposed by linguists, as follows:

Low vowels ... 10

Mid vowels ... 9

High vowels [i] [u] 8

... 7

... 6

Nasals [m] [n] ... 5

[s] 4

Voiced fricatives [v] ... 3

Unvoiced fricatives [f] [θ] [s] [∫] [x] 2

Voiced stops [b] [d] [g] 1

Unvoiced stops [p] [t] [k] 0.5

The values are relative only, and should not be thought of as numerical ratios. Openness of the vocal tract and voicing are the two variables that contribute to sonority levels. With this chart, it is possible to graph curves for any word, with the sonority peaks corresponding to the nuclei of syllables and the troughs to the syllable juncture points. Thus sonority can be used as a means to define the syllable. For the most part this system works, but accommodation must be made for some consonant clusters. A word such as straight will have two sonority peaks, because of the relative prominence of [s] over the ensuing [t]. It can also be seen from the chart that the most sonorous consonants (apart from [s]) are the ones that can function as syllable nuclei in spoken English: father / table / button. Musicians can also find the chart useful as a means of identifying which consonants require the most extra effort to optimize elocution and ensure intelligibility throughout words and phrases.


The present author has set out the guidelines for Italian syllabification in an earlier article, and thus will not repeat them here.8 Since this article concerns itself with the practical relevance for singers and pedagogues of knowing the guidelines for syllabification, it will suffice to concentrate on the two contexts in which the orthographic syllabification usually encountered in vocal texts can differ from the proper phonetic result.

1. The letter <s>, when initial in a cluster, should be transcribed as the coda of the preceding syllable, even in cases where the ensuing cluster would be permissible.

ba-sta ...

disposto ...

ma-estro ...

in-chiostro ...

Although the opposite may be encountered, both in vocal texts (ma-es-tro) and in transcription (...), it should be remembered that consonant clusters beginning with <s> always shorten the previous vowel. Since unstressed syllables always have short vowels, this is only of significance in stressed syllables.

2. Intervocalic <gli>, <gn>, <sc>, and <z> are to be geminated, both within a polysyllabic word and phrasally, even though this is not flagged in the syllabification of the text. Thus,

me-glio ...

so-gna-re ...

la-scia-re ...

a-za-le-a ...

Do-ni-zet-ti ...

a-zio-ne ...

The Italian language allows onset clusters not encountered in English, particularly with <s> + voiced consonant, as in sbaglio, sdengo, sguardo, slancio, smarrito, snello, sregolato, and svelte. Phrasally, the initial <s> can be considered a coda to the previous syllable, as in lo sdegno ... The word smuovere begins with two examples of consecutive phonemes not found in English, [zm] and [mw]. A similar challenge pertains in gli uomini, with its ... grouping.

Some consonant clusters foreign to English may begin medial syllables, but not initial syllables in a word, as in a-vria ... Such clusters would always be split at syllable junctures in English: sovereign ... / livery ... Care should be taken not to overly shorten the Italian vowel preceding such clusters.


The notion of syllable in German is intimately bound up with vowel length and the morphemic structure of the language, with its root stems and many affixes. The basic guidelines for orthographic syllabification are outlined here.

1. A single consonant will normally be in onset, as in sa-gen and be-ne-deit. This is the same as saying that long vowels are usually closed-a fundamental principle of German phonology. A long vowel will be in coda position in an open syllable, except with some monosyllabic words (Tag, Wahl, Weg).

2. The abrupt onset of voicing associated with a glottal stroke [?] before a vowel is a certain indicator of the beginning of a syllable (be-ach-ten). This often creates an exception to the first guideline ...

3. Devoiced consonants function as terminal markers (Mäd-chen, ab-ge-hen, weg-räumen; but Mä-del, Wege). In words with elided <e>s, such as and-rer, ed-ler, and seg-nen, the voiced consonant at the end of the root stem moves to coda position from the onset position it would occupy in the additional syllable, were the <e> present.

4. Double consonants are split between syllables, as in most other languages (bit-ter, Hasse, won-ne-voll). In the case of <ck>, the syllables are either written without a hyphen, or split before the cluster (we-cken), even though the archaic spelling equivalent was <kk>. In the case of <dt>, however, the cluster is split (Städ-te). In both cases the double consonants may be considered ambisyllabic.

5. Consonant clusters that represent more than one phoneme will usually divide between syllables, avoiding phonotactic impossibilities in the process (ach-ten, Sil-be, sin-ken).

6. Compound words have syllable breaks at the morpheme junctures (Mit-tag, Rück-kehr, Tanz-lied, bis-her, blau-äugig, gleich-falls).

7. Affixes usually define syllable boundaries at the morpheme juncture (an-neh-men, Be-frei-ung, berg-ab, Freund-schaft, Irr-tum, Lieb-chen, zer-stö-ren).

8. Digraphs and trigraphs begin ensuing syllables (lachen, zwi-schen), because they function phonetically as single phonemes (/x/, /∫/), as in rule 1. A single letter that represents two phonemes also begins the ensuing syllable, while the transcription will be split (Ni-xen ...

9. By convention, <st> is never split between syllables (be-ster, Gei-ster). The digraph <ng> is split, however (lan-ge), since it represents an ambisyllabic ...

10. Silent intervocalic <h> begins the ensuing syllable (fä-hig, flie-hen, Mü-hé).

11. Diphthongs are never divided (Ei-sen, Mau-er, Freu-de). While this can be considered a "universal" principle in syllabification for all languages, it nevertheless contrasts with vowel clusters that form separate syllables, as iati do in Italian (a-e-re, be-a-to, Lu-i-gi, mae-stro). No such hiatus between pure vowels exists in German, except across morpheme boundaries (be-achten), where a glottal stroke is essential. German does however have bisyllabic vowel clusters involving diphthongs, as in Bau-er, Schlei-er, and teu-er.

12. By convention, German orthography avoids separating a single syllabic vowel from an adjacent syllable. Thus, Abend and Reue, not A-bend and Reu-e. While important for typesetting, such words are regularly hyphenated in vocal literature, and considered the separate syllables that they are phonetically.

13. By convention, <ng> is split, even though it represents a single ... Hun-ger, sin-gen, Wan-ge. However, hung-rig, Sing-vögel, etc.

It will be seen that German orthographic syllabification usually does not follow the morpheme boundaries of root stems, preferring rather to begin ensuing syllables with the final consonant of the cluster, or the digraph. Thus,

With clusters of two or more consonants, it is usually the final phoneme that belongs to the next syllable. Thus, deut-sche splits before the final trigraph. One must be careful to dissociate the determination of word roots from syllabification, or one will often be led astray. The word bester has a short open [ε] because of the root best-, even though the word divides as be-ster. In other words, syllabification is often not a reliable guide to vowel length. Singers need to be aware that text syllabification in vocal music is not a substitute for root stem recognition in the determination of vowel length (and thus openness/closure).

The abundance of Fremdwörter (foreign borrowings) in German complicates the guidelines given above, and syllabification may follow patterns from the original language. As this is beyond the present article, the reader is referred to the relevant dictionary in the standard Duden series (Band 5: Fremdwörterbuch).

Spoken German features syllabic consonants, like English. The syllabic ... (Apfel), <m> (großem), and <n> (haben) of colloquial or rapid speech must be modified for singing by the insertion of a schwa ... before the consonant.

Accuracy in German lyric diction depends strongly upon a basic comprehension of the grammar of the language, the structure of sentences, and an ability to recognize the root stems and other morphemes upon which the vocabulary is derived. The German lexis is constructed in an agglutinative manner, with a disarming array of affixes and compound forms. The linear sequencing of morphemes is familiar from English, "antidisestablishmentarianism" being an extreme and overexploited example. Agglutinative German words like Volksliedbearbeitungen can be mind-benders without a basic ability to identify the constituent morphemes.9 Vowel color and vowel length in particular are closely connected to word structure. For the beginning student who may be able to recognize only a few of the root stems from which so many nouns, verbs, and adjectives are derived, this can be daunting challenge. For instance, the pronunciation of a word like Halbinsel ("peninsula") hinges upon recognition of its compound status, as Halb ("half") + Insel ("island"). Errors can occur because the incorrect syllabification Hal-bin-sel obeys the phonotactic rules of German, and because <lb> is not a coda cluster natural to English. The word does not obviously syllabify as Halb-insel unless the component words are recognized. Incorrect syllabification will result in ... a result perhaps unrecognizable to a German speaker, because of the absence of the medial glottal [?], crucial in identifying the word Insel. The lack of a devoiced [b] in Halb, and the existence of the word ... in the German lexis will also contribute to incomprehensibility. Many more such words can be cited. Alexander Pope warned us in 1709 that "a little learning is a dangerous thing," and German has many red herrings to derail the unwary. The word gestern ("yesterday") ... for instance, is a single morpheme, even though ge- ... and Stern [∫tεrn] are common morphemes. The reader may wish to consider how, in a similar manner, morpheme recognition implicates the transcription and pronunciation of the following words:






gestern / gestorben






zergehen / zurück


Anyone who has transcribed French vocal texts into IPA understands the particular nature of French transcription as it relates to the written language. French is a language of predominantly open syllables and, like English, has a particularly complex relationship between orthography and pronunciation. French IPA transcription provides a particularly apt and useful tool for clarifying pronunciation for lyric purposes. Among the principal languages of vocal literature, the distinction between orthographic and phonetic syllabification is most overtly demonstrated by French. Orthographic syllabification guidelines are amply covered in the literature on lyric diction, and need not be reiterated here.10 Emphasis instead will be given on suggested syllabification policies for transcribing vocal texts into IPA. The policy of transcribing all vocal texts with open V and C(CC)V syllables (except for the pronounced consonants at the end of poetic lines of text), is laudable from the point of view of eliminating the ever-present tendency to move too early to the next consonant. But, taken to an extreme, this policy results in transcriptions with awkward, illogical, and forbidden onset clusters, as in arbre ... and volupté ...

Singers possessing a fluent familiarity with the IPA often find that idiomatic pronunciation of French song and opera texts is easier to achieve directly from an IPA transcription, rather than from the original text. The reasons for this are twofold. First, IPA simplifies the orthographic complexity inherent in the written language, reducing to a single symbol the many possible spellings for a particular phoneme. Second, the syllabification principles of the written language operate in a manner that does not reinforce the basic principle of open syllables, and in many instances is quite in opposition to this essential phonological feature. Any randomly chosen text can serve to illustrate these points.11 Compare the texts below with their transcription:

Quel offre a-t-on faite à vos a-mis?


J'ai dans le coeur u-ne tris-tesse af-freu-se!


Every syllable is open in the above transcriptions, even though six words end visually in a consonant, and seven words (not always the same ones) when spoken in isolation would end phonetically in a consonant. A contrastive study of the syllable structure of English and French quickly reveals that French not only has a far greater percentage of open syllables, it also has far fewer examples of VC syllables (as in if, oeuf) than English, especially in style soutenu, where extra syllables are created by giving the e-muet syllabic status. The word difficile has all open syllables (CV/CV/CV/CV) in soutenu, but ends with a closed syllable (CV/CV/CVC) in speech. Diphthongs predominate over pure vowels in English open syllables, while French CV syllables are either pure vowels (trou) or vowel + semiconsonant (soleil). The tendency to delay consonants and semiconsonants at word junctures over to the next syllable is very strong in French, and is rendered as well in IPA transcription as it is poorly in orthography. There are limits to the consistent application of this policy in transcription, however. In the following cases, final consonants should remain and close a syllable in phonetic transcription, rather than leap forward to the next:

1. A final consonant at the end of a rhythmic group

2. Consonant clusters that would violate onset phonotactic rules if taken forward are split between syllables. Thus,

3. <x> [ks] is split between syllables.

vexer [vε]

examen [...]

examen [...]

4. <cc> [ks] and <gg> [g3] are split between syllables.

accident [ã]

suggérer [...]

5. Geminated double consonants are also split.

illusion [õ]

immortelle [...]

innombrable [in.nõ.bra.ble]

irrésolu [...]

But the following are not geminated:

ennemi []

ennuyé [...]

emmener [ã]

6. Intervocalic <y> [j] moves to the next syllable, or is split between syllables.

noyer []

fuyant [...]

It is best to think of the intervocalic <y> as a double <ii> (as indeed it was diachronically, when manuscript scribes wrote the double <ii> as <ij>, which is only a small step away from <y>).12 The respellings noi-ier and fui-iant then reveal their proper pronunciation more readily.

7. <s> + consonant(s). Warnant13 states that, within a word, <s> + a consonant splits, as in

It seems advisable however, for lyric diction purposes, to carry forward <s>s within words too ([...]), for the sake of legato and the avoidance of the English tendency to anticipate consonants too early. If the resultant cluster is not permissible in onset, the cluster then must be split, as in disgrâce [dis.gra: se], disjoint [...], and dislocation [...].14 Of course if the ensuing consonant is silent, the <s> carries forward, as in disciple [di. si. ple].

French mélodie texts have a particular way of being hyphenated in musical scores. This approach serves to preserve the meter of the original poem, and should not be construed as an "authentic" manner of breaking words into syllables from a purely linguistic point of view. The soutenu basis of metered French poetry demands a different set of syllabification rules than those designed to reflect the spoken tongue. French dictionaries generally syllabify neither headwords nor IPA transcriptions (when given), for a number of reasons, one of which is the dual systems involved in familier/soutenu. Warnant, since it is a dictionary of pronunciation only, adopts a policy of syllabification, but only for the transcriptions.


Juncture is a suprasegmental feature that refers to the boundary point between phonetic units, and can be thought of at the level of the word or the syllable. Juncture can be phonemic, in the sense that two distinct words or phrases can be identical phonetically and differ only in phonotactics. The most well known example in English is undoubtedly "I scream" [...] vs. "ice cream" [...], which constitutes a case of phrasal minimal pairs, if vowel length is ignored. On an analytic level, juncture is a difficult topic, but intuitively one can appreciate easily that the examples above differ primarily in the relative durations of the segments (i.e., the "placement" of the successive articulations). This has considerable consequence for singing, since musical rhythm dictates at least to some extent such placement, and the duration of vowels and consonants with respect to one another. A detailed look at this consideration must await a future article. It will suffice here to point out that juncture is an important factor in the comprehension of text in languages other than English as well. The following examples illustrate the importance of articulation timing for clarity of meaning.

English - white shoes / why choose

night rate / nitrate

that stuff / that's tough

an aim / a name

great ape / grey tape

at all / a tall

catch it / ( - - - - )

German - Eich-en ("oaks") / Ei-chen ("little egg")

den Bau erkennen / den Bauer kennen

Er dachte, sie wird kommen / Er, dachte sie, wird kommen

Standard IPA transcription as employed in lyric diction, segmentally oriented as it is, does not adequately account for all the modifications of intonation, stress, and length that is necessary in distinguishing such pairs. In French, the syllable-timed rhythm of the language, combined with the many lexical possibilities for small groups of articulations, yields many phrasal homophones that are indistinguishable phonetically, and must rely on context:

l'hôtel odieux / l'autel au Dieu

Much comical wordplay can result, as Alphonse Allais demonstrated in his numerous facetious holorime couplets:

Je dis, mettons, vers mes passages souterrains

Jeudi, mes tons verts, mais pas sages, sous tes reins.


Two competing criteria - morphological/syntactic and phonetic/acoustic-vie with one another in the "correct" determination of orthographic syllabification. In many words, one of these considerations prevails over the other. Some languages, like English, have more ambisyllabic consonants than others. This is a reflection of the ambiguous status of such consonants, and of the fact that the two criteria are often pulling in opposite directions. One of the primary uses of IPA transcription is its focus on exclusively phonetic/acoustic considerations when syllabifying. For singers, this has the effect of clarifying any pronunciation ambiguities that may arise from looking only at the printed text, and (especially in French) of reinforcing the correct timing of the phonetic events.

Predictability in the timing of any succession of sounds in singing is a crucial reality for vocal accompanists and for conductors who work with singers. Whether they are conscious of it or not, vocal accompanists wait until they expect the singer has completed any onset consonant or consonant cluster, placing an on-beat chord where they anticipate the phonation of the ensuing vowel will begin. This is true whether the onset consists of unvoiced or voiced consonants, or a combination of the two. Pulse in vocal music coincides with vowel phonation onset, and not with syllable onset, except in cases of VC, VCC, etc. Accompanists develop a sixth sense for the amount of time typically taken by various consonants and consonant clusters, and sensitive accompanying of foreign language repertory requires an awareness of the possible onset clusters of each language, and how they differ. Thus, the amount of time a singer spends on such clusters is a key ingredient of ensemble performing as well as of idiomatic text delivery and dramatic interpretation. If for no other reason, accompanists must develop a sense of how languages behave in terms of consonant/vowel interaction. Syllabification and juncture are intimately bound up with the manner in which the text meshes with the rhythm of the melodic line. The "rhythm" of the beginnings of each articulation in a phrase does not, and should not, equate precisely with the rhythm of the musical line. The ability to exploit the symbiosis between syllable and pulse for expressive effect is one sign of a compelling performer.


1. The terms syllabification and syllabication are both encountered and accepted. Which one uses depends on whether one syllabifies or syllabicates words-again, both forms accepted. This column employs the former.

2. If one adopts a more etymological approach, -extric- can be said to divide further into two morphemes, -ex- and-tric-, bringing the total to five.

3. The Chicago Manual of Style contains a thorough treatment of syllabification guidelines for American English, 13th ed. (Chicago & London: University of Chicago Press, 1982, 164-167).

4. The nucleus and coda together comprise the rhyme.

5. Some languages, like Japanese, acknowledge a grammatical unit known as the mora, similar to a syllable, but referring to agogic syllable "weight." A short syllable in Japanese has one mora, while a long syllable has two.

6. Prior to the acquisition of specific language skills, babies do exactly this, as they discover the myriad possibilities of sound combinations that are physiologically possible. The reader may wish to experiment with imitating infant babble (probably when no one is around), to verify the possibility of nonsense speech consisting entirely of unvoiced consonants, or entirely of voiced consonants, or a combination of the two-each entirely devoid of vowels.

7. This creates an issue on long held notes in Czech music, where [e] must be employed as an intrusive syllabic vowel for most of the duration, with [r] added at the end of the note. The same is done in English for words such as sudden, little, etc.

8. Leslie De'Ath, "Italian Vowel Clusters in Singing, Part I: Syllable, Vowel Length, and Stress," Journal of Singing 65, no. 3 (January/February 2009): 333-337.

9. The present example has seven: Volk / s / lied / be / arbeit / ung / en. Notice that the morphemic structure does not tally with the syllable count (also seven): Volks-lied-be-ar-bei-tung-en.

10. The reader is referred to David Adams, A Handbook of Diction for Singers, 2nd ed. (Oxford: Oxford University Press, 2008), 150-153, and to Thomas Grubb, Singing in French (New York: Schirmer Books, 1979), 12-15.

11. All transcriptions in this section will follow style soutenu guidelines regarding <r>, mute <e>, etc., rather than colloquial spoken French.

12. The letter <y> is called i grec for that reason.

13. Léon Warnant, Dictionnaire de la prononciation française dans sa norme actuelle, 4th ed. (Paris-Gembloux: Duculot), xcii-xciv, cvii-cviii.

14. The last example would involve a potential onset cluster [sl] that, while permissible in English, is found in French only in foreign borrowings.

Author affiliation:

Leslie De'Ath, Associate Editor

Author affiliation:

Leslie De'Ath is a Canadian vocal coach, pianist, and conductor. He is Professor in the Faculty of Music at Wilfrid Laurier University in Waterloo, Ontario, where he has taught since 1979. There he is Music Director of the Opera Program, teaches studio piano, and instructs performance literature and diction courses. His research interests, in addition to lyric diction and phonology, focus on unusual vocal and piano repertoire.

He is presently at work on research/recording projects of the complete piano music of Cyril Scott (on the Dutton Epoch label) and Florent Schmitt. Mr. De'Ath is the regular keyboard player with the Canadian Chamber Ensemble and the Kitchener-Waterloo Symphony, and has been a featured soloist with both groups on many occasions-in concert, on tour, and on CD. He has had a wide-ranging career as a vocal accompanist, chamber player, and vocal concert series administrator.

The use of this website is subject to the following Terms of Use