I am a Departmental Lecturer and Research Fellow of Phonetics at the University of Oxford. I was born in Greece and studied at the University of California, Los Angeles (BA, MA and PhD in Linguistics). I previously held a faculty position at the University of Ioannina. Supervisors: Sun-Ah Jun, UCLA, PhD Advisor, Anna Szabolcsi, UCLA (currently NYU), MA advisor, and Anna Szabolcsi, UCLA (currently NYU), BA Honors thesis advisor
This study examines how different phrasings in identical utterances in Greek are marked by pitch,... more This study examines how different phrasings in identical utterances in Greek are marked by pitch, pre-boundary lengthening, sandhi phenomena, and pausing, and the influence of speech rate on them. Eight speakers (4F, 4M) produced utterances in Greek containing structures of the type “noun (N) + clitic (C) + verb (V)” in two phrasings, [(N + C) (V)] (noun-final) and [(N) (C+V)] (clitic-final), in normal and fast speech rate. Results show that both phrasing and speech rate influenced different measures. Specifically, in both phrasings, pre-boundary lengthening affected the phrase-final word, the phrase-final syllable, as well as the phrase pre-final syllable. Significantly higher scaling and earlier alignment of the H tone was found in the clitic-final than in the noun-final phrasing. Finally, the clitic-final phrasing induced greater degree of sandhi effects. In fast rate, sentence, word and segment durations were shorter than in normal rate. Overall, results on the effect of phrasing agree with previous literature on Greek and other languages, while novel findings on the patterns are revealed by the combined effect of phrasing and speech rate.
10th International Conference on Speech Prosody 2020, 2020
Greek wh-questions are typically uttered with one of two tunes. Questions with the L*H L-H% tune ... more Greek wh-questions are typically uttered with one of two tunes. Questions with the L*H L-H% tune are interpreted as information-seeking; questions with the LH* L-L% tune may be interpreted as non-information seeking, carrying implicatures of a negative type. As the tunes are nuclear and differ in both pitch accent and boundary tone, we conducted two experiments to test the contribution of each to these pragmatic interpretations: participants (N = 66 in Exp1; N = 88 in Exp2) heard questions in which the stretch that included either the boundary tone (Exp1) or the pitch accent (Exp2) was high-pass filtered to remove F0 information. The participants were asked to bet €0-100 on the most likely of two utterances following the question, choosing between follow-ups that indicated the pragmatic intent of the question to be informationor noninformation-seeking. Bets on follow-ups indicating information-seeking intent were expected to be higher after questions with a L*H pitch accent (Exp1), or a H% boundary tone (Exp2). The results show that both pitch accents and boundary tones made pragmatic contributions and affected responses (i.e. strength of bets), thereby supporting the view that intonation meaning is compositional and nuclear tunes are not processed as a whole.
Asia Minor Greek (AMG) speakers cohabited with Turkish speakers for eight hundred years until the... more Asia Minor Greek (AMG) speakers cohabited with Turkish speakers for eight hundred years until the 1923 Lausanne Convention, which forced a two-way mass population exchange between Turkey and Greece and severed their everyday contact. We compare the intonation of the continuation rise tune in the speech of first-generation AMG speakers born in Turkey with three subsequent generations born in Greece. We examine how long contact effects in intonation persist after contact has ceased, through comparison of the f0 patterns in four generations of AMG speakers with those of their Athenian Greek- and Turkish-speaking contemporaries. The speech of the first-generation of AMG speakers exhibits two patterns in the f0 curve shape and time alignment of the continuation rises, one Athenian-like and one Turkish-like. Over subsequent generations use of the latter diminishes, while the Athenian pattern becomes more frequent, indicating intergenerational change.
H μeλέτη χρησιμοποιeί ακουστικά και ηλeκτροπαλατογραφικά δeδομένα για να eξeτάσeι την παραγωγή το... more H μeλέτη χρησιμοποιeί ακουστικά και ηλeκτροπαλατογραφικά δeδομένα για να eξeτάσeι την παραγωγή του eλληνικού /r/ σe συμφωνικά συμπλέγματα /Σr/ μe ηχηρά αποφρακτικά /b, d, g, v, ð, ɣ/. Τα αποτeλέσματα eπιβeβαιώνουν τη σύνθeτη παραγωγή του /r/, η οποία πeριλαμβάνeι ένα φωνηeντικό μέρος και μια σύντομη πeρίοδο μeρικής ή ολικής φραγής, καθώς και ποικιλία στη θέση και στον τρόπο άρθρωσης. Οι διαμορφωτές του φωνηeντικού μέρους έχουν πιο κeντρικές τιμές σe σχέση μe το φωνήeν της συλλαβής. Η σύγκριση των αποτeλeσμάτων μe προηγούμeνα δeδομένα σe συμπλέγματα μe άηχο αποφρακτικό δeίχνουν μeγαλύτeρη διάρκeια φραγής και μικρότeρη διάρκeια φωνηeντικού μέρους σe πeριβάλλον άηχου αποφρακτικού και σe πeριβάλλον έκκροτου σe σχέση μe τριβόμeνο σύμφωνο.
The sentence in (1) has been widely reported to have at least two interpretations (the most frequ... more The sentence in (1) has been widely reported to have at least two interpretations (the most frequently cited early reference for such examples is Jackendoff 1972), shown in (1a) and
The pragmatic interpretation of Greek wh-questions with different intonation was tested, by askin... more The pragmatic interpretation of Greek wh-questions with different intonation was tested, by asking participants to listen to questions and bet on two follow-up sentences offering alternative explanations on the question’s purpose (informationor noninformation-seeking). L*+H L-!H% and L+H* L-L% were used and crossed (L+H* L-!H% and L*+H LL%), giving rise to four experiment versions in a between-participant design. Responses from 190 Greek listeners supported previous analyses according to which L*+H L-!H% and L+H* L-L% lead to a preference for information-seeking vs. noninformation seeking interpretations respectively. Responses were affected by both the pitch accent and boundary tone, with the joint contribution being most evident in the “crossed” tunes (L+H* L-!H% and L*+H L-L%). These results also support the notion that accents and edge tones contribute independently to pragmatic meaning, while the successful application of betting as an experimental paradigm supports the idea th...
This study examined cross-dialectal differences on the perception of Greek vowels. Speakers of St... more This study examined cross-dialectal differences on the perception of Greek vowels. Speakers of Standard Modern Greek (SMG) and two dialectal areas (Crete, Kozani), all with five vowels in their systems, chose best exemplar locations (prototypes) for Greek vowels embedded in a carrier sentence spoken by a speaker of their dialect. The results showed that SMG, Cretan and Kozani vowels were well separated in the perceptual space. At the same time, there were dialect-induced differences in the positioning and distances between vowels as well as in the total space area covered by each dialect. The organisation of perceived vowel space therefore seems to be dialect-specific, a finding which is consistent with production studies examining the organisation of the acoustic vowel space.
Relatively little work exists regarding the impact of long-term language contact on intonation. I... more Relatively little work exists regarding the impact of long-term language contact on intonation. In this paper we give an overview of the project ‘Intonation and diachrony: a phonetic investigation of the effects of language contact on intonational patterns’. In this project we are investigating the intonational tunes of regional varieties of Greek whose speakers have a history of cohabitation with speakers of Venetian Italian (Cretan and Corfiot Greek) or Turkish (Asia Minor and Cypriot Greek). We compare the shape of the f0 contours of the contact dialects to the corresponding tunes in both Standard (Athenian) Greek and their respective donor languages. Standard Autosegmental–Metrical analysis is combined with statistical modelling of f0 curves using Functional Data Analysis. The robustness of contact effects over time is traceable through comparisons of speech corpora spanning a century.
We explore the morphophonology of the Greek glide in prevocalic position, where it surfaces eithe... more We explore the morphophonology of the Greek glide in prevocalic position, where it surfaces either as a palatal fricative or merges with a preceding consonant of the set [k, g, x, ɣ, n, l] to produce [c, ɟ, c, ʝ, ɲ, ʎ]. We offer empirical evidence for the glide’s dual nature, both as an allophone of /i/ and as a distinct phoneme and we demonstrate how an account based on morphological paradigms captures alternations between vocoids versus lack thereof. Our analysis also sheds light on palatalization. We show that the palatals are allophones of the velars, both before front vowels, as traditionally assumed, but also before back vowels, a fact so far unaccounted for. Moreover, we demonstrate that palatals are derived through two processes, simple and extreme palatalization.
This paper presents the results of a production experiment which examines the acoustic characteri... more This paper presents the results of a production experiment which examines the acoustic characteristics of the Greek rhotic sound in consonant clusters. The experimental results suggest that in this position the Greek rhotic has a different realization from the intervocalic rhotic: while in intervocalic position this sound is realized as a tap, in consonant clusters the tap closure is accompanied by a short vowellike transition between the tap closure and the second consonant of the cluster. The presence of the vowel-like transition was systematic, appearing in 79.5% of the tokens, and is discussed in reference to similar realizations found in many other languages.
In the extensive literature on the prosodic expression of Information Structure (IS) the notion o... more In the extensive literature on the prosodic expression of Information Structure (IS) the notion of contrast is typically coarse grained and subsumed under relational dichotomies like the theme-rheme or topic-focus, or as an inherent feature of focus, evoking a set of alternatives. This paper has two goals. First, we advocate for a more nuanced conception of contrast. This distinguishes between the “alternatives” based meaning of contrast on one hand and correction on the other, which is a more discourse-oriented meaning that encodes the speaker's assumptions about the hearer's beliefs. Second, we present experimental evidence that among the pragmatic types of contrast examined, only correction receives distinct prosodic marking, which cuts across the traditional IS topic-focus division and is realized in the same way in focus and topic constituents.
This paper presents the results of a production experiment which examines the acoustic characteri... more This paper presents the results of a production experiment which examines the acoustic characteristics of the Greek rhotic sound in consonant clusters. The experimental results suggest that in this position the Greek rhotic has a different realization from the intervocalic rhotic: while in intervocalic position this sound is realized as a tap, in consonant clusters the tap closure is accompanied by a short vowellike transition between the tap closure and the second consonant of the cluster. The presence of the vowel-like transition was systematic, appearing in 79.5% of the tokens, and is discussed in reference to similar realizations found in many other languages.
This study examines how different phrasings in identical utterances in Greek are marked by pitch,... more This study examines how different phrasings in identical utterances in Greek are marked by pitch, pre-boundary lengthening, sandhi phenomena, and pausing, and the influence of speech rate on them. Eight speakers (4F, 4M) produced utterances in Greek containing structures of the type “noun (N) + clitic (C) + verb (V)” in two phrasings, [(N + C) (V)] (noun-final) and [(N) (C+V)] (clitic-final), in normal and fast speech rate. Results show that both phrasing and speech rate influenced different measures. Specifically, in both phrasings, pre-boundary lengthening affected the phrase-final word, the phrase-final syllable, as well as the phrase pre-final syllable. Significantly higher scaling and earlier alignment of the H tone was found in the clitic-final than in the noun-final phrasing. Finally, the clitic-final phrasing induced greater degree of sandhi effects. In fast rate, sentence, word and segment durations were shorter than in normal rate. Overall, results on the effect of phrasing agree with previous literature on Greek and other languages, while novel findings on the patterns are revealed by the combined effect of phrasing and speech rate.
10th International Conference on Speech Prosody 2020, 2020
Greek wh-questions are typically uttered with one of two tunes. Questions with the L*H L-H% tune ... more Greek wh-questions are typically uttered with one of two tunes. Questions with the L*H L-H% tune are interpreted as information-seeking; questions with the LH* L-L% tune may be interpreted as non-information seeking, carrying implicatures of a negative type. As the tunes are nuclear and differ in both pitch accent and boundary tone, we conducted two experiments to test the contribution of each to these pragmatic interpretations: participants (N = 66 in Exp1; N = 88 in Exp2) heard questions in which the stretch that included either the boundary tone (Exp1) or the pitch accent (Exp2) was high-pass filtered to remove F0 information. The participants were asked to bet €0-100 on the most likely of two utterances following the question, choosing between follow-ups that indicated the pragmatic intent of the question to be informationor noninformation-seeking. Bets on follow-ups indicating information-seeking intent were expected to be higher after questions with a L*H pitch accent (Exp1), or a H% boundary tone (Exp2). The results show that both pitch accents and boundary tones made pragmatic contributions and affected responses (i.e. strength of bets), thereby supporting the view that intonation meaning is compositional and nuclear tunes are not processed as a whole.
Asia Minor Greek (AMG) speakers cohabited with Turkish speakers for eight hundred years until the... more Asia Minor Greek (AMG) speakers cohabited with Turkish speakers for eight hundred years until the 1923 Lausanne Convention, which forced a two-way mass population exchange between Turkey and Greece and severed their everyday contact. We compare the intonation of the continuation rise tune in the speech of first-generation AMG speakers born in Turkey with three subsequent generations born in Greece. We examine how long contact effects in intonation persist after contact has ceased, through comparison of the f0 patterns in four generations of AMG speakers with those of their Athenian Greek- and Turkish-speaking contemporaries. The speech of the first-generation of AMG speakers exhibits two patterns in the f0 curve shape and time alignment of the continuation rises, one Athenian-like and one Turkish-like. Over subsequent generations use of the latter diminishes, while the Athenian pattern becomes more frequent, indicating intergenerational change.
H μeλέτη χρησιμοποιeί ακουστικά και ηλeκτροπαλατογραφικά δeδομένα για να eξeτάσeι την παραγωγή το... more H μeλέτη χρησιμοποιeί ακουστικά και ηλeκτροπαλατογραφικά δeδομένα για να eξeτάσeι την παραγωγή του eλληνικού /r/ σe συμφωνικά συμπλέγματα /Σr/ μe ηχηρά αποφρακτικά /b, d, g, v, ð, ɣ/. Τα αποτeλέσματα eπιβeβαιώνουν τη σύνθeτη παραγωγή του /r/, η οποία πeριλαμβάνeι ένα φωνηeντικό μέρος και μια σύντομη πeρίοδο μeρικής ή ολικής φραγής, καθώς και ποικιλία στη θέση και στον τρόπο άρθρωσης. Οι διαμορφωτές του φωνηeντικού μέρους έχουν πιο κeντρικές τιμές σe σχέση μe το φωνήeν της συλλαβής. Η σύγκριση των αποτeλeσμάτων μe προηγούμeνα δeδομένα σe συμπλέγματα μe άηχο αποφρακτικό δeίχνουν μeγαλύτeρη διάρκeια φραγής και μικρότeρη διάρκeια φωνηeντικού μέρους σe πeριβάλλον άηχου αποφρακτικού και σe πeριβάλλον έκκροτου σe σχέση μe τριβόμeνο σύμφωνο.
The sentence in (1) has been widely reported to have at least two interpretations (the most frequ... more The sentence in (1) has been widely reported to have at least two interpretations (the most frequently cited early reference for such examples is Jackendoff 1972), shown in (1a) and
The pragmatic interpretation of Greek wh-questions with different intonation was tested, by askin... more The pragmatic interpretation of Greek wh-questions with different intonation was tested, by asking participants to listen to questions and bet on two follow-up sentences offering alternative explanations on the question’s purpose (informationor noninformation-seeking). L*+H L-!H% and L+H* L-L% were used and crossed (L+H* L-!H% and L*+H LL%), giving rise to four experiment versions in a between-participant design. Responses from 190 Greek listeners supported previous analyses according to which L*+H L-!H% and L+H* L-L% lead to a preference for information-seeking vs. noninformation seeking interpretations respectively. Responses were affected by both the pitch accent and boundary tone, with the joint contribution being most evident in the “crossed” tunes (L+H* L-!H% and L*+H L-L%). These results also support the notion that accents and edge tones contribute independently to pragmatic meaning, while the successful application of betting as an experimental paradigm supports the idea th...
This study examined cross-dialectal differences on the perception of Greek vowels. Speakers of St... more This study examined cross-dialectal differences on the perception of Greek vowels. Speakers of Standard Modern Greek (SMG) and two dialectal areas (Crete, Kozani), all with five vowels in their systems, chose best exemplar locations (prototypes) for Greek vowels embedded in a carrier sentence spoken by a speaker of their dialect. The results showed that SMG, Cretan and Kozani vowels were well separated in the perceptual space. At the same time, there were dialect-induced differences in the positioning and distances between vowels as well as in the total space area covered by each dialect. The organisation of perceived vowel space therefore seems to be dialect-specific, a finding which is consistent with production studies examining the organisation of the acoustic vowel space.
Relatively little work exists regarding the impact of long-term language contact on intonation. I... more Relatively little work exists regarding the impact of long-term language contact on intonation. In this paper we give an overview of the project ‘Intonation and diachrony: a phonetic investigation of the effects of language contact on intonational patterns’. In this project we are investigating the intonational tunes of regional varieties of Greek whose speakers have a history of cohabitation with speakers of Venetian Italian (Cretan and Corfiot Greek) or Turkish (Asia Minor and Cypriot Greek). We compare the shape of the f0 contours of the contact dialects to the corresponding tunes in both Standard (Athenian) Greek and their respective donor languages. Standard Autosegmental–Metrical analysis is combined with statistical modelling of f0 curves using Functional Data Analysis. The robustness of contact effects over time is traceable through comparisons of speech corpora spanning a century.
We explore the morphophonology of the Greek glide in prevocalic position, where it surfaces eithe... more We explore the morphophonology of the Greek glide in prevocalic position, where it surfaces either as a palatal fricative or merges with a preceding consonant of the set [k, g, x, ɣ, n, l] to produce [c, ɟ, c, ʝ, ɲ, ʎ]. We offer empirical evidence for the glide’s dual nature, both as an allophone of /i/ and as a distinct phoneme and we demonstrate how an account based on morphological paradigms captures alternations between vocoids versus lack thereof. Our analysis also sheds light on palatalization. We show that the palatals are allophones of the velars, both before front vowels, as traditionally assumed, but also before back vowels, a fact so far unaccounted for. Moreover, we demonstrate that palatals are derived through two processes, simple and extreme palatalization.
This paper presents the results of a production experiment which examines the acoustic characteri... more This paper presents the results of a production experiment which examines the acoustic characteristics of the Greek rhotic sound in consonant clusters. The experimental results suggest that in this position the Greek rhotic has a different realization from the intervocalic rhotic: while in intervocalic position this sound is realized as a tap, in consonant clusters the tap closure is accompanied by a short vowellike transition between the tap closure and the second consonant of the cluster. The presence of the vowel-like transition was systematic, appearing in 79.5% of the tokens, and is discussed in reference to similar realizations found in many other languages.
In the extensive literature on the prosodic expression of Information Structure (IS) the notion o... more In the extensive literature on the prosodic expression of Information Structure (IS) the notion of contrast is typically coarse grained and subsumed under relational dichotomies like the theme-rheme or topic-focus, or as an inherent feature of focus, evoking a set of alternatives. This paper has two goals. First, we advocate for a more nuanced conception of contrast. This distinguishes between the “alternatives” based meaning of contrast on one hand and correction on the other, which is a more discourse-oriented meaning that encodes the speaker's assumptions about the hearer's beliefs. Second, we present experimental evidence that among the pragmatic types of contrast examined, only correction receives distinct prosodic marking, which cuts across the traditional IS topic-focus division and is realized in the same way in focus and topic constituents.
This paper presents the results of a production experiment which examines the acoustic characteri... more This paper presents the results of a production experiment which examines the acoustic characteristics of the Greek rhotic sound in consonant clusters. The experimental results suggest that in this position the Greek rhotic has a different realization from the intervocalic rhotic: while in intervocalic position this sound is realized as a tap, in consonant clusters the tap closure is accompanied by a short vowellike transition between the tap closure and the second consonant of the cluster. The presence of the vowel-like transition was systematic, appearing in 79.5% of the tokens, and is discussed in reference to similar realizations found in many other languages.
Uploads
Papers by Mary Baltazani