This thesis suggests that the range of vocabulary in an individual’s lexicon has an influence on in their assessment of nonword wordlikeness. The study included thirteen Spanish-English bilinguals who participated in a language dominance questionnaire, standardized assessments of Spanish and English vocabulary knowledge, and Spanish and English wordlikeness judgment tasks. Resulting data demonstrated moderate correlations between vocabulary knowledge and performance on nonword wordlikeness judgement tasks in Spanish and English. Participants with larger lexicons appeared more tolerant of less probable nonwords, those with low phonotactic probability, while those with smaller lexicons were less accepting of nonwords with low phonotactic probability. The results suggest that an individual’s processing of low probability phonological constituents is influenced by the diversity and complexity of their linguistic knowledge and specifically, their vocabulary acquisition.

This thesis addresses the problem of automatic speech recognition, specifically, how to transform an acoustic waveform into a string of words or phonemes. A preliminary chapter gives linguistic information potentially useful in automatic speech recognition. This is followed by a description of the Wave Analysis Laboratory (WAL), a rule-based system which detects features in speech and was designed as the acoustic front end of a speech recognition system. Temporal reasoning as used in WAL rules is examined. The use of WAL in recognizing one particular class of speech sounds, the nasal consonants, is described in detail. The remainder of the thesis looks at the statistical analysis of samples of spontaneous speech. An orthographic transcription of a large sample of spontaneous speech is automatically translated into phonemes. Tables of the frequencies of word initial and word final phoneme clusters are constructed to illustrate some of the phonotactic constraints of the language. Statistical data is used to assign phonemes to phonotactic classes. These classes are unlike the acoustic classes, although there is a general distinction between the vowels, the consonants and the word boundary. A way of measuring the phonetic balance of a sample of speech is described. This can be used as a means of ranking potential test samples in terms of how well they represent the language. A phoneme n-gram model is used to measure the entropy of the language. The broad acoustic encoding output from WAL is used with this language model to reconstruct a small test sample. "Branching" a simpler alternative to perplexity is introduced and found to give similar results to perplexity. Finally, the drop in branching is calculated as knowledge of various sets of acoustic classes is considered. In the work described in this thesis the main contributions made to automatic speech recognition and the study of speech are in the development of the Wave Analysis Laboratory and in the analysis of speech from a phonotactic point of view. The phoneme cluster frequencies provide new information on spoken language, as do the phonotactic classes. The measures of phonetic balance and branching provide additional tools for use in the development of speech recognition systems.

Economy growth rate is an index of a nation¡¦s gross productivity. China, Russia and Thailand are a few nations whose economy growth rates exceed the global average. In the recent years, the rapid development in China, including the enhanced relation with Taiwan, has made her the member of the BRICS, the top five emerging countries in the world. Russia has been playing an important role in the international society during the past decades. She is not only the member of the G8, the group of eight major industrial nations, but also her language, Russian, is one of the six official languages in the United Nations. According to the statistics of the Taiwan Funds, Russia and Thailand are the top two countries in their investment growth. Thailand, located in the middle of the Southeast Peninsular, together with Malaysia and Philippines, are the three founding members of the ASEAN 10, the Association of Ten Southeast Asian Nations. Due to the industrial and household needs, Taiwan has offered job opportunities to foreign labors from the Southeast countries. Therefore, it is our objective to design a trilingual speech recognition system for Chinese, Russian and Thai to meet the needs of language learning and household living. This system utilizes 404 Chinese, 611 Russian and 123 Thai common mono-syllables, selected from their pronunciation rules, as the major speech training and recognition methodology. Mel-frequency cepstral coefficients, linear predicted cepstral coefficients, and hidden Markov model are used as the two syllable feature models and the recognition model respectively. Under the AMD 2.2 GHz Athlon XP 2800+ personal computer and Ubuntu 9.04 operating system environment, the correct phrase recognition rates of 88.87%, 84.31% and 87.58% can be reached using phonotactical rules for the 82,000 Chinese, 31,883 Russian and 3,809 Thai phrase database respectively. Furthermore, a trilingual language-speech recognition system for 300 common words, composed of 100 words from each language, is developed. A 98.66% correct language-phrase recognition rate can be obtained.

History, culture and economy constitute the foundation of language. Mandarin Chinese is our native language, spoken by over 1.2 billion people. Its population is ranked number one in the world. In the recent years, the emerging China not only possesses market and labor forces, but also develops the Chinese culture circle in Asia. British history and American politics make English the most influential language in the 20th century. Vietnam has been under the profound influence of Chinese culture. The reformed and opened economy in the past decade brought her tremendous foreign investments, including those from Taiwan. It is our objective to establish a trilingual system for travel, living and speech learning. This thesis investigates the design and implementation strategies for a trilingual speech recognition system of Chinese, English and Vietnamese. It utilizes the speech features of 404 Chinese, 925 English and 154 Vietnamese mono-syllables as the major training and recognition methodology. Mel-frequency cepstral coefficients, linear predicted cepstral coefficients, and hidden Markov model are used as the two syllable feature models and the recognition model respectively. Under the AMD XP 2800+ personal computer and Ubuntu 9.04 operating system environment, the correct rates of 88.16%, 82.74% and 87.45% can be reached using phonotactical rules for the 82,000 Chinese, 30,795 English and 3,300 Vietnamese phrase database respectively. The computation for each system can be completed within 2 seconds. Furthermore, a trilingual language-speech recognition system for 300 common words, composed of 100 words from each language, is developed. A 98% correct language-phrase recognition rate can be obtained with the computation time less than 2 seconds.

Nonword repetition tasks were originally devised to assess the efficiency of the phonological loop (Baddeley & Hitch, 1974), a component of the working memory system, where verbal information is temporarily stored and translated to support activities like phonological processing during early word-recognition (Snowling, 1981; Wagner et al., 2003), speech production (McCarthy & Warrington, 1984), and articulation (Watkins, Dronkers, & Vargha-Khadem, 2002; Yoss & Darley, 1974). From a practical perspective, there is a significant need for a systematically-designed Spanish nonword repetition measure that is equivalent to currently-available English measures. For this study, a database of nonwords that considered phonotactic and phonological properties of Spanish was devised. In a preliminary study, Spanish-speaking adults provided wordlikeness judgments about a large set of candidate nonwords. A subset of the rated nonwords was used in the development of a Spanish nonword repetition measure. The aim of the main experiment was to explore the contributions of participant factors (age, gender, and vocabulary knowledge) and item factors (word length, stress pattern, and wordlikeness) to Spanish repetition performance in this group of Spanish speaking, English language learning children. From a theoretical perspective, this investigation allowed a first observation of how experience with listening to and producing Spanish words influences the acquisition of Spanish-specific phonological patterns. A total of 68 children, ages four to six years with varying degrees of Spanish language knowledge participated in this study. Results revealed significant age and word length effects. However, stress pattern did not exert significant effects on repetition performance, which is not completely consistent with previous literature. That is, participants repeated nonwords from both the more frequent and the less frequent stress pattern with similar accuracy. Wordlikeness, a previously uninvestigated variable in nonword repetition was found to affect repetition accuracy. For all participants, nonwords rated as high in wordlikeness were more accurately repeated than were nonwords with low wordlikeness ratings. Findings of the study are discussed in terms of how they relate to working memory and usage-based models of phonological learning. Finally, the clinical relevance of nonword repetition in the assessment of coarse- and fine-grained mappings of phonological knowledge is suggested.

No description available.

The correct use of morphological suffixes in obligatory contexts reflects linguistic knowledge and competence of speakers. Grammatical knowledge is acquired during a child’s period of primary language acquisition, and may be partial or incomplete due to normal linguistic variation found during acquisition, due to a child’s level of progression through typical chronological development, or due to the presence of language disorders, like specific language impairment (SLI). In the current study, we ask whether characteristics of verbs make it more or less likely that children will correctly use an inflectional morpheme. The morphemes of interest in the current study were third person singular –s (3S) and past tense –ed (ED). Data for analysis were taken from a database of spontaneous language samples collected from 40 children (20 with SLI and 20 developing typically; Hoover, Storkel, & Rice, 2012). Spontaneous language samples were analyzed for the presence or absence of each morpheme in obligatory contexts. For each word item, the uninflected base word was additionally analyzed for a number of phonological and lexical variables. After comparing children with SLI to typically developing peers group differences emerged with respect to the effect of phonological and lexical variables. Moreover, different variables were determined to predict the 3S and ED morphemes. The results are discussed highlighting relevant theoretical and clinical implications.

Animals detect, recognize and localize relevant objects in noisy, multi-source environments. Female crickets locate potential mates in choruses of simultaneously calling males using acoustic signals, a behaviour termed phonotaxis. The mechanisms underlying cricket phonotaxis are now understood across multiple levels: biophysical, neurobiological and behavioural. Phonotaxis has, however, rarely been tested in the complex real-world acoustic environments and no attempts have been made to predict acoustic orientation behaviour in these conditions despite our extensive understanding of its underlying mechanisms. In this thesis, I first characterized the acoustic environments faced by female crickets of the species Plebeiogryllus guttiventris in the field. Phonotaxis behaviour of females was then characterized under laboratory conditions using two sound sources. The data obtained were used to develop a simulation that predicted this behaviour. The predictions of the simulation were then tested against the phonotaxis behaviour of females in realistic, multi-source conditions in the field. My field studies of male behaviour showed that males of this species produced complex and variable songs in choruses where multiple males called simultaneously. The acoustic ranges of males in these choruses overlapped extensively and females performing phonotaxis in such choruses would hear multiple males simultaneously. The acoustic interactions of simultaneously calling males were also characterized for their timing relationships with each other and the changes they made to the temporal patterns of their songs. Males did not either synchronise or alternate their chirps, however they made changes to the temporal patterns of song in a way that is likely to make them more attractive to females. I then characterized the closed-loop walking phonotaxis behaviour of P. guttiventris females in the presence of two active sound sources playing conspecific song. Both the baseline and relative SPLs of the two speakers were systematically varied and female phonotactic paths were obtained. Females were found to preferentially approach louder songs. Several aspects of this behaviour were characterized, in particular orientation ability and motor behaviour under varied conditions of stimulus intensity. A stochastic simulation of closed-loop walking phonotaxis behaviour was developed using both current understanding of field cricket physiology and my data on closed-loop walking phonotaxis. The simulation was demonstrated to both qualitatively and quantitatively recapture female behaviour. It was also able to qualitatively recapture female behaviour in two previously published classical experiments in which the hearing of female crickets was disrupted. Female phonotaxis was then tested under real-world multi-source conditions. The behaviour of real females was compared to the predictions of the simulation. The simulation was found to recapture both female preference and phonotactic path forms at the population level. To my knowledge, this is the first study to both examine and successfully predict phonotaxis behaviour in complex real-world acoustic conditions.

Examensarbetet bygger på tidigare forskning som visat att två egenskaper hos ord, hög/låg fonotaktisk sannolikhet och hög/låg grannskapstäthet, kan påverka ordinlärning hos vuxna och barn. I examens­arbetet upprepas ett ordexperiment som tidigare genomförts i USA och Neder­länderna. Syftet var att, genom en experimentell design, undersöka påverkan av fonotaktisk sannolikhet och grannskapstäthet på ordinlärningen hos 21 elever i åk 3 med typisk språk­ut­veckling. Sambandet mellan elevernas ordinlärning och vokabulär, fonologisk medvetenhet och/eller verbalt arbetsminne undersöktes också. Varje elev genomförde ett ordexperiment i en sagokontext med inlärning av 16 nonord som representerar svenska substantiv. De verbalt och visuellt förmedlade orden har skapats utifrån den svenska fonotaxen. Eleven mötte orden upprepade gånger och skulle genom sagans olika episoder lära in och reproducera dessa. Denna studies resultat är i linje med tidigare resultat från USA och Nederländerna. Eleverna genom­förde också ett test av ordförrådets bredd, PPVT-IV, fonologisk medvetenhet och verbalt arbets­minne, NEPSY-II.  Statistisk analys (ANOVA) visade att repetition förbättrade inlärning­en. En korrelationsanalys visade ett starkt positivt samband mellan antalet inlärda ord och fonologisk medvetenhet. Då det fanns en effekt av hög/låg fonotaktisk sannolikhet i interaktion med hög/ låg grannskapstäthet hos elever med typisk språkutveckling, kan det i vidare forskning vara av intresse att undersöka om denna effekt även gäller för elever med språkstörning och/eller dyslexi. / This student thesis is based on former research which claims that certain characteristics of words, high/low phonotactic probability and high/low neighbourhood density, may affect word learning in adults and children. In this thesis a word experiment, formerly used in the USA and the Netherlands, was replicated. The aim was through experimental design, examine the impact of phonotactic probability and neighbourhood density on word learning in 21 pupils in grade 3 with typical language development. Correlations between pupils’ word learning and vocabulary, phonological awareness and/or verbal short-term memory were also examined. Each pupil performed a word experiment through a story context where 16 pseudo words representing Swedish nouns were learned. The verbally and visually presented words have been created using Swedish phonotactics. The student encountered the words repeatedly and was tasked to during the different episodes learn and produce the words. The result of this study is in line with former studies in the USA and the Netherlands. The pupils also performed tests of vocabulary, PPVT-IV, phonological awareness and verbal short-term memory, NEPSY-II. Statistical analyses (ANOVA) suggested that repeated training improves learning. Correlation analysis indicated a strong positive correlation between number of words learned and phonological awareness. Since there was an effect of high/low phonotactic probability combined with high/low neighbourhood density in pupils with typical language development, future research may examine this effect in pupils with developmental language disorder and/or dyslexia.

<p>Istraţivanje je imalo za cilj da provjeri u<br />kojoj mjeri se na&scaron; kognitivni sistem moţe<br />osloniti na fonotaktiĉke informacije, tj.<br />moguće/dozvoljene kombinacije fonema/<br />grafema, u zadacima automatske percepcije i<br />produkcije rijeĉi u jezicima sa bogatom<br />infleksionom morfologijom.<br />Da bi se dobio odgovor na to pitanje,<br />sprovedene su tri studije. U prvoj studiji, uz<br />pomoć ma&scaron;ina sa vektorima podr&scaron;ke (SVM),<br />obavljena je diskriminacija promjenljivih<br />vrsta rijeĉi. U drugoj studiji, produkcija<br />infleksionih oblika rijeĉi izvedena je<br />pomoću uĉenja zasnovanog na memoriji<br />(MBL). Na osnovu rezultata iz druge studije,<br />izveden je eksperiment u kojem se traţila<br />potvrda kognitivne vjerodostojnosti modela i<br />kori&scaron;ćenih informacija.<br />Diskriminacija promjenljivih vrsta rijeĉi<br />obavljena je na osnovu dozvoljenih sekvenci<br />dva i tri grafema/fonema (tzv. bigrama i<br />trigrama), ĉije su frekvencije javljanja<br />unutar pojedinaĉnih gramatiĉkih tipova<br />izraĉunate u zavisnosti od njihovog poloţaja<br />u rijeĉima: na poĉetku, na kraju, unutar<br />rijeĉi, svi zajedno. Maksimalna taĉnost se<br />kretala oko 95% i dobijena je na svim<br />bigramima, uz pomoć RBF jezgrene<br />funkcije. Ovako visok procenat taĉne<br />diskriminacije ukazuje da postoje<br />karakteristiĉne distribucije bigrama za<br />razliĉite vrste promjenljivih rijeĉi. S druge<br />strane, najmanje informativnim su se<br />pokazali bigrami na kraju i na poĉetku rijeĉi.<br />MBL model iskori&scaron;ćen je u zadatku<br />automatske infleksione produkcije, tako &scaron;to<br />je za zadatu rijeĉ, na osnovu fonotaktiĉkih<br />informacija iz posljednja ĉetiri sloga,<br />generisan traţeni infleksioni oblik. Na<br />uzorku od 89024 promjenljivih rijeĉi uzetih<br />iz Frekvencijskog reĉnika dnevne &scaron;tampe<br />srpskog jezika, koristeći metod izostavljanja<br />jednog primjera i konstantu veliĉinu skupa<br />susjeda (k = 7), ostvarena je taĉnost oko<br />92%. Identifikovano je nekoliko faktora koji<br />su uticali na ovu taĉnost, kao &scaron;to su: vrsta<br />rijeĉi, gramatiĉki tip, naĉin tvorbe i broj<br />primjera u okviru jednog gramatiĉkog tipa,<br />broju izuzetaka, broj fonolo&scaron;kih alternacija<br />itd.<br />U istraţivanju na subjektima, u zadatku<br />leksiĉke odluke, za rijeĉi koje je MBL<br />pogre&scaron;no obradio utvrĊeno je duţe vrijeme<br />obrade. Ovo ukazuje na kognitivnu<br />vjerodostojnost uĉenja zasnovanog na<br />memoriji. Osim toga, potvrĊena je i<br />kognitivna vjerodostojnost fonotaktiĉkih<br />informacija, ovaj put u zadatku<br />razumijevanja jezika.<br />Sveukupno, nalazi dobijeni u ove tri studije<br />govore u prilog teze o znaĉajnoj ulozi<br />fonotaktiĉkih informacija u percepciji i<br />produkciji morfolo&scaron;ki sloţenih rijeĉi.<br />Rezultati, takoĊe, ukazuju na potrebu da se<br />ove informacije uzmu u obzir kada se<br />diskutuje pojavljivanje većih jeziĉkih<br />jedinica i obrazaca.</p> / <p>The study was aimed at testing the extent to<br />which our cognitive system can rely on<br />phonotactic information, i.e., possible/<br />permissible combinations of phonemes/<br />graphemes, in the tasks of automatic<br />processing and production of words in<br />languages with rich inflectional<br />morphology.<br />In order to obtain the answer to this<br />question, three studies have been conducted.<br />In the first study, by applying the support<br />vector machines (SVM) the discrimination<br />of part of speech (PoS) with more than one<br />possible meaning (i.e., ambiguous PoS) was<br />performed. In the second study, the<br />production of inflected word forms was<br />done with memory based learning (MBL).<br />Based on the results from the second study,<br />a behavioral experiment was conducted as<br />the third study, to test cognitive plausibility<br />of the MBL performance.<br />The discrimination of ambiguous PoS was<br />performed using permissible sequences of<br />two and three characters/sounds (i.e.,<br />bigrams and trigrams), whose frequency of<br />occurrence within individual grammatical<br />types was calculated depending on their<br />position in a word: at the beginning, at the<br />end, and irrespective of position in a word.<br />Maximum accuracy achieved was<br />approximatelly 95%. It was obtained when<br />bigrams irrespective of position in a word<br />were used. SVM model used RBF kernel<br />function. Such high accuracy suggests that<br />brigrams&#39; probability distribution is<br />informative about the types of flective<br />words. Interestingly, the least informative<br />were bigrams at the end and at the beginning<br />of words.<br />The MBL model was used in the task of<br />automatic production of inflected forms,<br />utilizingphonotactic information from the<br />last four syllables. In a sample of 89024<br />flective words, taken from the Frequency<br />dictionary of Serbian language (daily press),<br />achieved accuracy was 92%. For this result<br />the MBL used leave<br />-one<br />-out method and nearest neighborhood size of 7 (k = 7). We</p><p>identified several factors that have<br />contributed to the accuracy; in particular,<br />part of speech, grammatical type, formation<br />method and number of examples within one<br />grammatical type, number of exceptions, the<br />number of phonological alternations, etc.<br />The visual lexical decision experiment<br />revealed that words that the MBL model<br />produced incorrectly also induced elongated<br />reaction time latencies. Thus, we concluded<br />that the MBL model might be cognitively<br />plausibile. In addition, we reconfirmed<br />informativeness of phonotactic information,<br />this time in human conmprehension task.<br />Overall, findings from three undertaken<br />studies are in favor of phonotactic<br />information for both processing and<br />production of morphologically complex<br />words. Results also suggest a necessity of<br />taking into account this information when<br />discussing emergence of larger units and<br />language patterns.</p>