At the XIIth International Congress of Phonetic Sciences held in Aix-en-Provence (19–24 August 1991), Professor Peter Ladefoged was presented with a Gold Medal, jointly offered by the International Organizing Committee of the Congress, the International Phonetic Association, and the International Society of Phonetic Sciences, in recognition of his contribution to the subject of phonetics over the forty years of his career to date. Professor Mario Rossi, as the President of the International Organizing Committee of the Congress, asked me to say a few words at the banquet in Peter Ladefoged's honor. This article is intended to give an expanded version of those comments a more permanent form, and to bring an appreciation of his outstanding contribution to a wider audience.

This paper examines the speech of two functionally misarticulating Jordanian children and illustrates the variability of misarticulation patterns as well as the failure of markedness to account for them. The presentation is very similar to work by Gierut describing misarticulating English-learning children (Gierut 1984, 1985), but in this case the children are learning Arabic. The term ‘functional misarticulation’ is typically used to describe the speech of speakers whose chronic articulatory errors cannot be attributed to any obvious organic disorder such as hearing impairment or cleft palate (Ingram 1976, Compton 1970). The basic assumption of much of the work done on speech misarticulations is that children's knowledge is identical to that of the ambient speech community (Compton 1975, Lorentz 1976). Within this framework, misarticulators are viewed as a homogenous group. Any discrepancy between the misarticulators' system and the ambient system is described as a ‘process’ (Shriberg & Kwiatkowski 1982). However, the many diverse phonological rules that have been posited to derive the misarticulators' surface structure from underlying structure make it difficult to arrive at any but gross commonalities across functional misarticulation systems (Ingram 1976). Such an assumption is a clear misrepresentation of the apparent differences across the misarticulation systems (Dinnsen, Elbert & Weismer 1980).

Formant measurements show sex and age differences in the formant patterns of a single vowel category. Comparisons of the formant frequency values of men, women and children indicate low, middle and high values, respectively (Chiba & Kajiyama 1941, Potter & Steinberg 1950, Peterson & Barney 1952). The differences are found for all vowel categories, and they have generally been interpreted as a consequence of different vocal tract size.

The results of a spectrographic study of the Kabardian vowel system confirm reports in the literature that phonetic vowel quality in Kabardian is determined by the preceding consonant. The data further show that there are three distinct phonetic categories along the F1 axis in the vowel space, and that the contextually determined variation in vowel coloring generally respects this division. Durationally, a correspondence was found between vowel height and duration. While the durational differences were found to be comparable to those reported in languages with contrastive vowel length, the differences were interpreted as a phonetic strategy to enhance the perceptual salience of the vertical vowel system.

Many languages in South East Asia have voiceless nasal consonants that contrast with their voiced counterparts. What has not been reported previously is that there are two distinct types of voiceless nasals. We will begin by considering the more well-known type, found in languages such as Burmese. These voiceless nasals are usually said to have an open glottis for most of the articulation, but some voicing for the period just before the stricture is broken (Ladefoged 1971: 11). In this view they have two parts. The first is necessary to distinguish them from their voiced counterparts. The second distinguishes one voiceless nasal from another, by making the place of articulation more apparent; the voiced offglide from the nasal into the vowel has formant transitions that are characteristic of each place of articulation. Dantsuji has shown that, in addition, the voiced portion of the voiceless nasals in Burmese “includes significant cues which can distinguish points of articulation” (Dantsuji 1986: 10). Despite the small voiced portion, in most phonological treatments of these sounds, a voiceless nasal consonant is considered not as a sequence of two items, a voiceless segment followed by a voiced segment, but as a single voiceless unit with a low level phonetic rule inserting the voicing towards the end. Burmese is a textbook example, with forms as shown in Table 1 (from Maddieson 1984; also in Ladefoged 1993: 282).

It is generally accepted in the descriptive phonetic literature that most variations in pulmonic pressure observed during the production of speeech are passive reactions to changing glottal and supra-glottal resistance (Ladefoged 1968, Ohala et al. 1979, Rothenberg 1968). Active short-term positive changes in lung volume are only found with heavily stressed syllables (Ladefoged 1968, Ohala et al. 1979 ), and in rare cases with with particular consonants such as LuGanda geminates (Ladefoged 1971), Korean fortis stops (but see Rothenberg 1968) and possibly with certain Jingpho segments (Ladefoged 1968). It is assumed therefore that pulmonic speech sounds are normally produced against a constant mean background pressure by movement of air that occurs with decreasing volume of the lungs (Ladefoged 1967, 1968) and with constant respiratory muscle innervation (Rothenberg 1968). The two reported exceptions to this are the pulmonic suction [1] found in the ritual language Damin (Catford 1977) and the recently reported two ingressive fricatives in an ideolect of a Tsou dialect (an Austronesian language) (Fuller 1990). In all other cases, pulmonic initiation is pressure.

The previous issue of the Journal contained a discussion by Peter Ladefoged about interpreting the information in a speech waveform (JIPA, 21, 32–34), noting that examination of waveforms displays is becoming commonplace with the easy availability of personal computer tools for digitizing and editing. As Ladefoged noted:

“Several aspects of sounds are clearly distinguishable from the waveforms of a phrase. Stop closures are very evident, as are differences between voiced sounds which have repetitive waveforms and voiceless sounds which do not. Differences in amplitude can be used to distinguish high frequency, high intensity sibilants from lower intensity non-sibilant fricatives; and nasals and laterals usually have smaller amplitudes than the louder adjacent vowels. An expanded view of the waveform allows us to see intervals between peaks in the damped wave of a voiced sound, and thus to calculate the frequency of the first formant. Nasals can often be distinguished from vowels in these expanded waveforms, not only by their smaller amplitudes but also by the less clear formant structure.”

Yes-no questions, when unmarked morphologically and syntactically, are marked intonationally. The intonation of this kind of question is generally characterized either by a terminal rise or a higher overall fundamental frequency (F0). Ultan (1978) found that 38 of a sample of 53 languages use the first pattern, and 18 the second. Even in tone languages, these two patterns are not unusual. For example, yes-no questions in Hausa have a final jump to an extra high level regardless of the nature of the final tones (Miller & Tench 1980); in Mandarin both the top line and the bottom line delimiting the pitch range are raised in question intonation (Shen 1990). However, the existence of a downward intonation as an alternative, for instance in English (Bolinger 1989) and French (Fonagy & Berard 1973) has been reported. Although a number of languages have been shown to have a fall for this kind of question, it remains uncertain just how the property of being a question is conveyed. This paper attempts to investigate this issue through an acoustical analysis and a perceptual test of Changsha Chinese, a 6-tone Chinese dialect of the Xiang group, spoken in Hunan. This dialect is selected, because (1) no tone sandhi occurs in this dialect, facilitating the interpretation of the F0 contours in the acoustical analysis; and (2) no empirical study of F0 in this dialect has been previously conducted. While the majority of empirical works on intonation use data collected from read speech, this study undertakes the exploration of natural speech.

In the IPA chart which was the outcome of the IPA 1989 Kiel Convention, the glottal fricatives [h, ɦ] were included, as they were in previous charts. Unlike Kloster-Jensen (JIPA 1991: 42), I believe we were correct to leave [h, ɦ] as they were, in the space for glottal fricatives. In order to justify his belief that the [h] is not a glottal fricative, Kloster-Jensen recruits two good referees, Ladefoged (1962) and Brücke (1876). Other phoneticians also share the same belief, that the [h] is not a glottal fricative (for example, Pike 1943: 140, O'Connor 1973: 143–144), but many others do believe that [h] is really a glottal fricative. Here I want to give two good reasons why we should consider [h, ɦ] glottal fricatives, and leave them as they are in the IPA chart. The first reason is a general one, and is based on aerodynamic reasoning. The second is based on observations of productions of [h] in Hebrew, in Arabic and in Finnish.

The spectrogram in the previous issue of the Journal was of the phrase “It's good to make a guess.” A number of readers interpreted the last four words correctly, but only a few managed to work out the whole sentence. This issue is concerned with a different kind of phonetic data.

Until recently, phoneticians have not spent a great deal of time analyzing waveforms; but the advent of computer systems that make it easy for personal computers of all kinds to record and display sounds has led to this becoming a more common practice. Several aspects of sounds are clearly distinguishable from the waveforms of a phrase. Stop closures are very evident, as are differences between voiced sounds which have repetitive waveforms and voiceless sounds which do not. Differences in amplitude can be used to distinguish high frequency, high intensity sibilants from lower intensity non-sibilant fricatives; and nasals and laterals usually have smaller amplitudes than the louder adjacent vowels. An expanded view of the waveform allows us to see intervals between peaks in the damped wave of a voiced sound, and thus to calculate the frequency of the first formant. Nasals can often be distinguished from vowels in these expanded waveforms, not only by their smaller amplitudes but also by the less clear formant structure.

Esling (1990) reports on the computer coding of the IPA as amended to 1989. The coding system consists of three-digit numbers, divided into groups for consonants, vowels, and diacritics by the first digit. The set of coding numbers in the series 6nn was reserved for the symbols for voice quality and pathological speech listed in the Extensions to the IPA drawn up by the Group working on these symbols at the Kiel IPA Convention.

It is painful to admit neglectfulness, but I see no way around it in the following case. And better late than never. I was inattentive as I sent in my vote of approval to the outcome of the IPA 1989 Kiel Convention.

The two notes in JIPA 20(2) by Kohler (1990) and Ladefoged (1990) concerning the phonemic status of present-day Standard German ç] and [x] are one of many pieces of evidence that distributional (‘taxonomic’) phonemics has happily survived the thirty-year war with Generative Phonology and its offspring. But it is common knowledge among linguists that even half a century after Bloch's (1948) classic paper there is still no fixed and exhaustive set of postulates for phonemic analysis. Such questions as partial overlapping or neutralization or—especially important—‘grammatical prerequisites’ (Pike 1947, 1952) are still open issues, and it is quite probable that, at least for the last-named problem, there is no single, universal solution. In fact, it may very well be that languages differ inherently in this respect, and that for some of them the decision cannot be made in categorial terms. In nonextreme cases there may be at least two different solutions, each valid within its respective framework, one based on the assumption of the analytical primacy of grammatical (or part-grammatical) analysis, and the other on the reverse assumption of pure phonetic distribution. But even with juşt one of these alternatives, one given phonetic-environmental description may lead to a number of different solutions, as exemplified with particular conspicuity by Łobacz (1973). Admitting the alternative of primacy of morphemic analysis vs. pure phonetic distribution, she demonstrated that 504 (sic) different phonemic interpretations of one kind of Standard Polish are possible.