Prosodic System Research Articles

Njerep a postcard from the edge

Kunama has been reported by different scholars as having two or three tones, downstep (or not), contrastive length of both consonants and vowels, and lexical stress. Despite this range of reported phenomena, little in-depth research into the prosodic system of Kunama has been undertaken. The aim of the present study is to report such a detailed investigation and to establish on a solid footing basic aspects of the tonal system of Kunama. The work reported is preliminary in the sense that its scope is limited: we present phonological and phonetic evidence for the existence of three level tones, which can combine to form a number of contour tones. This is followed by discussion of tonal phenomena in the noun phrase. No evidence for lexical stress is found.

Trisyllabic shortening in English: past and present

Trisyllabic shortening or TSS is one of the most controversial processes in the history of English. Time after time, claims have been made about quantity variations, attributing them either to trisyllabic shortening, closed syllable shortening, or other mechanisms. Our examination of the nature of TSS in the history of English leads us to conclude that it differs from closed syllable shortening, which is syllable based, and that the preference for a maximal foot has remained the underlying incentive for maintaining vowel quantity variations throughout the centuries. However, the prosodic system has undergone dramatic changes and many features of TSS in the older and modern stages are not the same. Older TSS affected mostly inflected native words, while in Modern English, TSS causes alternations in derivationally related words with Romance suffixes. Interacting with open syllable lengthening, older TSS led to quantity alternations in inflectional paradigms which were later levelled out. Romance loans, both suffixed and nonsuffixed forms, were borrowed in their entirety and constrained by the prosodic structure of the language. Only later, when these words came to be derivationally related, were quantity alternations observable with TSS operating as a constraint dictated by the prosodic structure of the modern language. Thus, throughout the history of English, TSS has served the same purpose: it led to the preferred prosodic structure of the word.

Footed tones and tonal feet: rhythmic constituency in a pitch- accent language

While prosodic systems are typically either tonal or metrical, composite systems including both tone and metrical structure present yet another, albeit less frequent, possibility (Prince 1983, Hayes 1995). The focus of this paper is a prosodic system of the latter, more complex variety, evidenced in the Neo-stokavian dialect of Serbian or Croatian (henceforth NS; see Ivic 1958, 1985, Zec 1993). Although previously analysed as a pitch-accent system with no role allotted to the rhythmic structure (Browne & McCawley 1965, Ivic 1965, 1976, Halle 1971, Kenstowicz 1974, Lehiste & Ivic 1986, Inkelas & Zec 1988), on closer inspection NS discloses important resemblances with stress systems. Metrical structure is an independent agent in NS, as will be demonstrated here, and the entire prosodic system is characterised by a rich interplay between the tonal and metrical components.Co-presence of tone and foot structure may in principle result in several types of interfaces between these two components. Cases documented in the literature are of two types: those in which the distribution of tone is constrained by metrical constituency, and those in which constraints operate in the opposite direction. third type of case has also been documented. In Japanese, as analysed in Poser (1984, 1990), the accentual system is tonal in nature while various templatic phenomena call for an inventory of iambic feet, yet the two systems are independent of each other and do not interact. The former type is exemplified by a number of languages extensively discussed in the literature: Creek (Haas 1977, Halle & Vergnaud 1987), Kirundi (Goldsmith 1987, Goldsmith & Sabimana 1989, Hayes 1995), Seneca (Prince 1983: 82–86), Winnebago (Susman 1943, Miner 1979, 1981, 1989, Halle & Vergnaud 1987, Hayes 1995 and the references therein) and Ancient Greek (Golston 1989, Sauzet 1989). The other type of unilateral interactions, with the rhythmic structure dependent upon the distribution of tone, is instantiated, for example, by Golin (Bunn & Bunn 1970, Hayes 1995).In contrast, NS presents a case of bilateral interaction between tone and foot structure: tone exerts influence on the repertory of feet, and foot structure, in turn, constrains the distribution of tone. Although previously analysed as an instantiation of unilateral interaction, on a par with Golin (Inkelas & Zec 1988, Hayes 1995), new evidence to be presented here clearly disputes this position. This case is of immediate theoretical relevance for establishing the range of possible foot inventories. The resulting inventory is richer than in cases generally reported in the theoretical literature (Prince 1990, Mester 1994, Hayes 1995), and as such suggests a possible direction in which foot inventories may expand.

Stress in ASL: empirical evidence and linguistic issues.

The study of signed languages provides an opportunity to identify those characteristics of language that are universal and to investigate the effect of production modality (signed vs. spoken) on the grammar. Over time, American Sign Language (ASL) has accommodated itself to the production and perception requirements of the manual/visual modality, resulting in a prosodic system that is comparable in function to spoken languages but different in means of expression. The present focus is on phrasal prominence in ASL. I review the marking of stress and phrase boundaries in ASL, and discuss prominence assignment at the phrasal level, with brief mention of lexical stress. At the kinematic level, there is a modality effect in marking of linguistic prominence but no modality effect with respect to marking phrase position. Of significance is the fact that ASL lacks phrasal prominence plasticity, that is the ability to move prominence to mark focus in a sentence location other than phrase final. I review the typological implications of how ASL handles prominence as compared to other languages.

A metrics of advertising: Sound figures and processing responses

Sound figures are departures from language norms that use rhythm (sound intensity) and rhyme (sound identity) to establish metrical patterns. These figures are characteristic of poetry, borrowed by print and electronic advertisements to elicit consumer effects. The purpose of this article is to develop a metrics of advertising by adapting formalist literary criticism to advertising text and by generating hypotheses that link sound patterns to consumer responses. The article grounds advertising metrics in syllable‐stress theory, presents the method of metrical analysis, and demonstrates the use of the prosodic notational system by means of advertising examples. The methodological system enables hypothesis generation in consumer processing areas such as affective responses, dual entry routes, attention, short‐term memory, and long‐term memory.

Foot Structure and Word Stress in Middle Dutch

Whereas word stress in modern Dutch is subject to continuous debate (cf. Kager 1989, Trommelen and Zonneveld 1989, Booij 1995), word stress in the older stages of the language is seldom discussed. One of the few exceptions is Zonneveld's (1992, 1993a) investigation of word stress in the Middle Dutch (MNL) text 'Het Leven van Sinte Lutgart' [Saint Lutgart's Life]. Zonneveld's work points out several interesting issues. His main conclusions are, however, challenged in this paper. Zonneveld's first conclusion is that not much has changed in the last 700 years with respect to word stress and that the stress system in 'Lutgart' is very similar to that of modern Dutch. In this paper I argue that the prosodic system of 'Lutgart' resembles that of the old and middle stages of the other West Germanic (WGmc) languages. It still has a Germanic foot, and moreover, Open Syllable Lengthening (OSL) has not yet run its full course. This latter claim challenges the assumption of all MNL grammars, namely that OSL took place in Old Dutch (ONL) and was a fact in MNL. The only evidence given for this assumption comes from rhyme. However, rhyme in 'Lutgart' does not motivate this assumption. Zonneveld's second conclusion is that the prosodic variation found in 'Lutgart' is largely due to the status of the schwa as a stress attractor. This confirms his synchronic analysis of modern Dutch, which also assumes the stress attracting nature of schwa (Trommelen and Zonneveld 1989, Kager and Zonneveld 1986). I will argue that the variation is not due to the nature of schwa per se, but is due to prosodic preferences at play at the time of 'Lutgart'. Variation is found only in certain environments, which partly coincide with those where English showed Trisyllabic Shortening (TSS) (Lahiri and Fikkert, 1996). Dutch did not have TSS but chose another way of dealing with less favoured prosodic structures, as we will see below.

Optimal syllables are not always optimal: A prosodic structure of Yukaghir

The object of this paper is to investigate the prosodic system of Yukaghir from the perspective of Optimality Theory. The analysis reveals that the prosodic word in Yukaghir exhibits continuous left-to-right parsing into bimoraic monosyllabic feet. As a result, the surface occurrence of light monomoraic syllables is prohibited, and various strategies apply to fix the output. From a general point of view this article aims to show that a grammar may include a requirement for a foot to be aligned with both edges of the same syllable. When the constraints on foot structure play a critical role in determining the surface shape of the word, this requirement forces input-unfaithful heavy closed syllables.

Cross-linguistic evidence for the acquisition of accent by the onset of word use

An investigation of the acoustic correlates of stress or accent in infant vocalizations was undertaken using five English- and five French-learning participants at two developmental points, the onset of word use (10–13 months) and late in the single-word period (14–19 months). A data base of 555 disyllabic utterances (words and babble) was digitized from naturalistic recordings and quantified acoustically by indices of duration, amplitude, and fundamental frequency (F0). Cluster analysis as well as the results of perceptual judgments (i.e., both a bottom-up and a top-down approach) revealed a developmental trend toward the adult prosodic system. In addition, acoustic comparisons of French and American infants at both developmental points revealed differences in their use of amplitude, F0, and duration to accent a syllable. The differences were traceable to the adult patterns for the respective adult languages. These results suggest that infants’ perceptual attunement to prosody in the prelinguistic period manifests itself in the beginnings of adultlike production as early as the onset of word use. [Work supported in part by the UK Economic and Social Research Council.]

A Linguistic History of Italian

Part 1 Introduction: perspectives and aims the emergence of Italian chronology and historical sources a note on phonetic transcription some concepts in linguistic change. Part 2 History of the sound system: the prosodic system vowels consonants major vowel changes the glides [j] and [w], and their effects palatalizing effects of the front vowels consonantal weakening consonant lengthening, and syllable structure syllable-final weakening the word-final consonants dissimilation and metathesis the phonetic sources of inflectional -e and -i a note on learned phonology. Part 3 Structural evolution of nouns, adjectives and verbs: structure of words infectional morphology of the noun and adjective invariance and inflectionalessness number in the noun and adjective gender adjectives the demonstratives the articles the verb the pronouns the prepositions derivational word formation. Part 4 History of sentence structure: verb - subject agreement interrogation negation conjunctions comparison verb complementation aspect and tense mood. Part 5 Variation in modern Italian: types of variation the dialects structural influences of the dialects on popular Italian emergence of neostandard Italian? Italian abroad.

Prosodic overlay and headless feet in Yidip

In this paper we reexamine the prosodic system of Yidiji (Australian; Dixon 1977a, b; 1991) as a case supporting an approach to stress assignment in which the construction of bounded feet and the implementation of metrical heads within them occur as separate events (the SEPARABILITY hypothesis; Halle & Vergnaud 1987a,b (= H&V)). Specifically, we argue that a layer of feet assigned early in Yidiji phonological derivations is unheaded and insensitive to weight distinctions, and plays a role in allomorphy. Metrical heads are later assigned to these constituents by independent rules operating under quantity-sensitive principles. In other words, Yidiji employs an additive metrical system in which stressable heads are overlaid on existing foot structures (prosodic overlay). It is the interaction between the outputs of these operations which produces the flexible rhythmic structure for which Yidiji is known. This flexibility is exemplified by the forms in (la) and (lb) (from Dixon 1977a: 40–41), where stress falls on even-numbered and odd-numbered syllables respectively, counted from the left.

Rules of stress assignment in Québec French: Evidence from perceptual data

ABSTRACTThis article seeks to contribute to the analysis of the prosodic system of Québec French in a variationist framework. However, major problems have to be resolved before one can give a comprehensive view of this system, one of these concerning stress assignment. In order to examine the validity of the traditional view claimed for French by most linguists, which states that stress is always superposed on the final syllable of lexical words, different perceptual tests were devised and carried out with two groups of students. The results of these tests show the effect of structural factors on the perception of stressed syllables in Québec French and demonstrate that this canonic rule does not always hold for varieties of Québec French. On the basis of these results, a subset of stress rules is then proposed and illustrated with examples of words tested in the perceptual tests. Counterexamples to these rules are thereafter discussed, and a reconsideration of the principles of syllabification that are usually taken for granted for French is proposed in light of variable rules.

Finding and using stress accent in spoken language understanding

Stress accent is a feature of the prosodic system in English that functions syntagmatically to organize an utterance into shorter sections. Knowledge about the location of stress accents in the segmental stream can help a speaker‐independent spoken language understanding system prune the search space, e.g., for the correct lexical item or the correct pragmatic interpretation. Unlike automatic systems, human listeners are able to locate accents because they know what accented versus unaccented syllables sound like and because they can normalize for such things as speaker and speaking rate. Here, the results of a study designed to locate accents in continuous speech are reported. An attempt is made to use both speech knowledge and normalization by identifying salient characteristics of accent in training data along with accent characteristics of the test utterance. Accented and unaccented syllables are separated using selected clustering algorithms on a feature space of energy and duration measures. In addition, some specific ways in which the output of the accent finder can be used by other components of the system (e.g., lexical access, pragmatics) are suggested. [Work partially supported by DARPA/ISTO.]

Initial prominence and accent in Seoul Korean prosody

This paper investigates two questions about the prosodic system of Seoul Korean—whether there is a word‐initial prominence, and whether a concurrent F0 peak is best seen as an optional pitch accent or as a boundary tone associated with some prosodic level. It analyzes the utterances of two speakers of the Seoul dialect. The durations of syllables and vowels placed in various positions within a sentence are compared. The two intonational analyses are tested by noting the position of the F0 peak, determining how consistently the peak appears, and noting the effects of pitch range variation. Results suggest: (1) that initial syllables of a word are consistently longer than medial syllables; (2) the position of the initial F0 peak is closely bound to the initial syllable of the utterance, but not of the word; and (3) the initial F0 peak appears in all tokens, except relatively short or fast tokens. This lack is shown to be due to tonal undershoot of a following low tone, as is apparent from strong and continuous relationships between height of pitch fall and rise and the temporal distance between the F0 peaks. These results suggest there is a word initial prominence in Seoul prosody, distinct from initial tones. The tones are best analyzed as boundary markers for some larger prosodic level, which become associated with initial syllables.

The Main Features of South Slavic Accentology

In this paper I explore the gross features of accentological development in Serbo-Croatian [SC], Slovene [Sn] and Bulgarian [Bg]. This survey will perforce be just a special case of the general history of Slavic accentology, but with particular South Slavic features manifesting themselves, especially in the later developmental stages. I begin my account in late Common Slavic [CS] at a point when the three main accentological paradigms for nouns, adjectives and verbs have been established: the barytonic (root-stressed) pattern; the final-columnar [FC] pattern (stress on the first syllable of the desinence); and the mobile pattern, where the stress alternates in a fixed pattern between the first and final syllables of the inflectional forms. Just how these three basic paradigms evolved is a fascinating and critical part of the complete story, but, unfortunateIy, lies for the most part beyond the scope of this discussion. I need say only that barytonic and FC words contain derived roots or stems, while mobile words are essentially nonderived forms without any obvious suffixes. As derived forms, barytonic and FC words are found with lengthened-grade roots and normal-grade roots, respectively; and may represent chronologically different levels of derivation. In Table I are displayed the late CS paradigms for typical South Slavic words, and the paradigms for the same words in the modem languages. For Late CS, note particularly the position of stress and the length of the accented syllable; and note that at the stage chosen here there are no syllables with rising intonation-all accented syllables, both long and short, have unmarked falling intonation. For the modem languages, note the degree of agreement with the CS paradigms: J The most spectacular phonological event in the history of the Slavic languages, with far-reaching side-effects on the prosodic system, was the loss of the jers, producing the neo-acute accents. It is reasonable to assume that the jers were lost in three stages: I) internally, 2) final when unstressed, 3) final when stressed. Weak internal jers appeared among a few fern. compound postverbals, collective nouns, and derived adjectives. The first category is nicely exemplified by *0 = SUp +a ~ Sn 6spa, SC ospa, 2 where the jer was lost and the stress drawn back to the resulting long syllable. Collective nouns with (joccurred among fern. and neut. nouns: *soux + iJ +a ~ Sn susa, SC susa, Bg sits a ; *storg +iJ+a ~ Sn strata, SC strata, Bg strata; *trrn +iJ+e ~ Sn tfnje, SC tfnje, Bg trane. In a more advanced stage of the original formula we have, e.g., *susiJa, where the jer disappeared and the stress was retracted to yield the neo-acute (rising) accent on the resulting long syllable. The falling accent in the SC forms was due to a later shift from a marked rising to an unmarked falling stress. The same suffix, -iJ-, was used with neut. o-stems to form verbal nouns on the past pass. pcpJe -t-/-nstem. Here we frequently observe the same kind of retraction: SC pietenje tresenje Sn pletcnje trescnje, Bg pletene tresene; but such formations must have arisen over a considerable time span, because we have final stress in most disyllabic forms and a few others: SC branje bdenje b(ce uzece, Sn branje bedcnje bftje vzitje. Much leveling has occurred in Bg: brane bdene cetene plene krlene. It is very likely that we find the same retraction among the pronominalized forms of derived adjectives with FC stress. A good case in point here is the adjective *kort-ii +k-J-jl, in all its inflected forms. Here the pretonic medial jer was lost, and the stress retracted to produce neo-acute accents on the preceding long syllable: Sn kratki, SC kratk f, Bg

The use of primary sentence stress by normal, aphasic, and autistic children

Primary sentence stress is an important aspect of the English prosodic system. Its adequate use is a prerequisite in the development of normal intonation patterns. This study examined the use of primary sentence stress in autistic children with mean length of utterance (MLU) scores between 1.9 and 4.1 morphemes. Normal and aphasic subjects at similar MLU levels served as contrast groups. The primary sentence task required that the subjects verbally respond to a request for information and provide a description of a play situation. Toy manipulation was used to elicit the desired responses. Listener judgment served as the basis for analyzing results. Although all subjects were able to perform the task, differences were seen in the number of correct responses and in the pattern of stress misassignment. These results are at variance with a prediction of stress placement on grammatical grounds. An explanation is offered, based on pragmatic considerations and cognitive developmental trends in young children.

Speech perception as a multilevel processing system

This is the report of an experiment that emphasized how variability in voice characteristics will affect a listener's ability to process sentences. Consideration is also given to the description of a possible model for the processing of continuous speech. As part of the experiment, use was made of a number of speakers, each of whom recorded the same set of sentences. The recorded sentences were used in order to form a pair of experimental tapes. On one tape, sentences were arranged randomly such that the listener could predict neither the content of a sentence nor the voice of the speaker who had produced it. On a second tape, sentences were arranged in blocks by speakers such that there was uncertainty as to content but not with respect to voice. Four groups of subjects listened to these tapes, with each tape being processed under a low and high condition of background noise. In interpreting the data, emphasis was placed on the idea that the speech signal contains a variety of characteristics or features and that these features are processed by three interacting subsystems. One is a prosodic system, responsible for the segmentation of continuous speech into sentences, phrases, and words. It attempts to establish a context within which a second system, responsible for the processing of words and syllables, can operate. However, this pair of systems is speaker-dependent in that it makes use of features that need to be adjusted on the basis of an assessment of voice characteristics. Thus, the model also provides for the inclusion of a third system, responsible for the assessment of voice characteristics. In effect, this third subsystem is responsible for “training” the two primary systems in order to normalize for differences in individual voice characteristics. We had predicted that performance would be uniformly more accurate in the blocked condition than in the unblocked condition since one type of uncertainty would have been eliminated. The results confirm this hypothesis only in part since there is an interaction with variations in backgroud noise. However, a plausible explanation is offered for these findings and for a variety of related results.

DE EXEGEET ALS RESTAURATEUR EN INTERPREET EEN VERHANDELING OVER DE BIJBELSE POËETICA MET PS. 121 ALS EXEMPEL. I.

Summary Ancient Hebrew poetics is an important subject for the exegete who respects the intricate unity of form and meaning. An insight in the prosodical form of a poem is essential for the interpretation of that poem. This essay tries to describe the ancient Hebrew prosodic system and deals with the determination of concrete prosodical structures (in this number of Bijdragen), with the relationship between prosodical structure and meaning and with the value of prosodical structure in the interpretation (in Bijdragen 44(1983) number 4). It is an interim report of my graduation study (see note 1). The first task in the exegesis of a biblical poem is to determine ('restore') the prosodical structure of that poem, because that has not or poorly been transmitted by tradition (§1.1). On the basis of two hundred years of research, but without discussion or much reference, I give a systematic description of the ancient Hebrew prosodic system (§1.2). This hypothesis contains three elements: (1) the prosodical structure as an hierarchical structure of text units, from the text as a whole via the higher text unit, the strophe and the line to the smallest prosodical unit, the colon (§1.21); (2) the prosodical rules, which determine the size of the units on each level (§1.22); (3) the system of balance, which combines two or more units of one level toa new unit on a higher level. Balance cannot be restricted to the level of the line but plays a role on all text levels. It can be brought about by all three textual constituents, rhythm, syntax and semantics, in different combinations (§1.23). Paragraph 1.3 gives an example of prosodical analysis, the determination of the prosodical structure of Psalm 121 (synthesis in §1.31). It starts with a systematic stock-taking of textual data (rhythmical, syntactical and semantical; §1.32) and continues with an examination of these data in which the prosodical structure on all levels is determined according to the above mentioned prosodic system (§1.33; cola and lines §1.331; strophes and higher text units §1.332). As for the strophes, their internal coherence is discussed (§1.3321), the divisions between them (§1.3322) and the regularity of the prosodical structure as a whole (§1.3323). Paragraph 1.4 contains an evaluation of the description of the prosodic system in the light of the exemplary analysis of Psalm 121. It pays attention to the working of the prosodic system. A further specification and extension of prosodical rules is asked for. The balance system has to be analysed on the different text levels in search for the working of the textual constituents separately and in relation with each other. Paragraph 1.5 criticizes the way in which known translations present the prosodical structure of biblical poems.

DE EXEGEET ALS RESTAURATEUR EN INTERPREET EEN VERHANDELING OVER DE BIJBELSE POËTICA MET PS. 121 ALS EXEMPEL. II

Summary The first part of this essay appeared in Bijdragen 44(1983)234–261. There I gave an elementary description of the ancient Hebrew prosodic system and dealed with the determination of concrete prosodical structures. In this second part I will draw the attention to the relationship between prosodical structure and meaning and to the value of the prosodical structure in the interpretation. Paragraphe 2.1 contains a discussion on the status of the prosodical structure. We may not take for granted that the meaning of a poem develops according to the prosodical structure of that poem. In many poems there is a small or big divergency between the prosodical structure and the meaning. (Some examples are given.) But meaning does only exist in mediating unity with form. Therefore, meaning which takes a course, ‘independent’ from the prosodical structure, needs its own structural data or its own structure to be supported. In this way we can explain the coexistence of two structures in one poem: first, the pros...

The Effect of Visualizing Speech Rhythms on Reading Comprehension and Fluency

The numerous difficulties associated with learning to read have often been contrasted with relatively few difficulties experienced when learning to understand speech. Gleitman and Rozin (1977, p. 3) suggested that visual system may be disadvantaged because the eye is not biologically adapted to However, eye does not have benefit of information given to ear through prosodic system of language. This automatically puts reader at a disadvantage: information which is readily available to listener is missing when listener becomes a reader. Sentences printed on this page do not have any visual pattern that could be compared with auditory patterns that they would have if they were spoken, such patterns as those produced by temporal sequence of short and long syllable durations, accented and unaccented levels of stress, and rise and fall of intonation. The prosodic system of language can be described by characteristics of duration, intensity (or stress) and intonation. Every syllable has either a short or long duration. Fraisse (cf. Ehrlich 1960) maintained that ratio of short to long is 2:1. The variation of syllable intensity seems to alternate between stressed and unstressed (Allen 1972). Syllables with strongest or primary stress are considered as accent points or unaccented syllables. These variables have been found to play two roles in speech perception (Darwin, 1975). The role which has been more thoroughly researched is in enabling temporal prediction of accents. Research has shown that intervals between accent points are perceived to be equidistant (Stroud 1955, Martin 1972, Allen 1975) can be temporally anticipated. Some evidence for this is given by Shields, McHugh & Martin (1974), Martin (1975), Cutler & Foss (1977) and Cutler & Fodor (1979) who found that reaction times to accented syllable targets in speech were quicker than to unaccented syllable targets. A temporal pattern is set up by sequence of long and short syllable durations and regular recurrence of accented and unaccented syllables. These patterns are referred to as having a rhythm (MacDougall, 1903). When a sentence does not meet expectations determined by rhythm (for example in experimental splicing) comprehension of speech is affected.

The Phonemics of Modern Icelandic

0.1. Background. The past three decades have seen a number of monographs and articles on the phonetics of Icelandic, but little has yet been done with its phonemics.l A first attempt was made by Kemp Malone in a paper read to the Linguistic Society in 1950, but this was drastically amended in a note he published three years later in LANGUAGE.2 There is a vigorous dogmatism in Malone's approach to phonemics, unhampered by those ifs and buts which strew the discourse of more cautious scholars. Although he claims to have taken account of phonemic work done since 1923, when he published his Phonology of Modern Icelandic, there are no references to indicate the scope of his reading, and few traces in his work of the ideas advanced by other scholars. His only statement of principle is this: 'I have tried first of all to classify each isolable speech-sound as an allophone of some phoneme.' He provides minimal pairs for each phoneme and a brief statement of its distribution. But we do not learn just how he isolated these speech sounds and overlooked others, or what criteria permitted him to classify the sounds with one phoneme and not another. Neither is there any description of the larger units into which the phonemes enter, or any consideration of alternative analyses such as we have been taught to expect in recent phonemic work. That alternatives are possible becomes apparent in his reconsideration of the prosodic system, which permitted him to reduce the number of phonemes at one blow from 61 to 39. This change suggests that there is room for further consideration of Icelandic phonemics. It will be the purpose of this paper to explore various possible analyses of Modern Icelandic as a contribution to what has been called the 'indeterminacy' of phonemic systems and the 'convertibility' of different descriptions. We shall also take up some of the problems involved in finding the structure or structures of a language.3