Journal of Linguistics and Language Teaching
Volume 4 (2013) Issue 1 (PDF)
pp. 49 - 75
Non-Native Speech Intelligibility of English Learners of Spanish:
the Impact of Gender, Aptitude and Motivation
Ángel Osle Ezquerra (London, UK)
This study offers an assessment of the non-native speech intelligibility of a group of English learners of Spanish in connected speech. Sixty evaluators, native speakers of Peninsular Spanish, transcribed different speech samples belonging to a group of 20 English learners of Spanish. The Spanish version of the Harvard Psychoacoustic Sentences served to elicit the different speech samples. Correlation analyses, using Pearson r, as well as the t-test, served to examine the relationship between three variables (gender, aptitude and motivation) and students’ intelligibility scores. Aptitude was measured through an oral mimicry task and a working memory capacity battery of tests, while a motivation questionnaire was used to study participants’ degree of motivation towards pronunciation accuracy. Results revealed that both aptitude and motivation were moderately correlated with intelligibility scores, while gender did not seem to exert any influence on degree of speech intelligibility.
Key words: speech intelligibility, aptitude, motivation, individual differences, L2 phonological acquisition.
Este estudio examina la inteligibilidad del habla de un grupo de aprendientes ingleses de español. Sesenta evaluadores, hablanbtes nativos de español peninsular, procedieron a la transcripción de muestras de habla pertenecientes a un grupo de 20 aprendientes ingleses de lengua española. Utilizamos las Harvard Psychoacoustic Sentences para elicitar las diversas muestras de habla. Hicimos uso, como herramienta estadística, del coeficiente Pearson r y la prueba t para estudiar la posible relación entre tres variables (sexo, aptitud y motivación) y las puntuaciones de los participantes en la prueba de inteligibilidad. Evaluamos la variable aptitud gracias a una prueba para medir la capacidad de imitación oral y a una batería de tests para medir la capacidad de la memoria de trabajo. Utilizamos un cuestionario para medir la motivaciòn de nuestros participantes con respecto a la importancia de la pronunciación. Los resultados obtenidos muestran que tanto la aptitud como la motivación se encuentran moderadamente correlacionadas con el grado de inteligibilidad del habla. Por otro lado, el sexo de los participantes no ejerció ninguna influencia sobre su nivel de inteligibilidad.
Palabras clave: inteligibilidad del habla, aptitud, motivación, diferencias individuales, adquisición fonológica de una L2.
It is widely accepted within communicative language teaching that intelligibility represents a sufficient goal for pronunciation instruction (e.g. Derwing 2008, Munro 2008, Munro & Derwing 2011). Attainment of a native accent, considered as unrealistic for those supporting the existence of a critical or sensitive period in Second Language Acquisition (SLA), has been replaced by the ‘intelligibility principle’ which posits that communicative effectiveness is a sufficient goal in itself (Levis 2005). As Kenworthy notes,
very few teachers today would claim that a pronunciation that is indistinguishable from that of a native speaker is necessary or even desirable for their learners. Instead, it is generally accepted that intelligibility is the most sensible goal. Kenworthy (1996: 13)
Furthermore, research has shown that ‘accentedness’ and intelligibility are not necessarily correlated (Derwing & Munro 2009). In other words, a heavy accent does not always imply a reduction in speech intelligibility.
The acceptance of the intelligibility principle seems to suggest that pronunciation instruction should be firmly grounded on empirical research relating, first and foremost, to speech intelligibility. However, the truth is that there is a great deal of anecdotal and theoretical commentary regarding the teaching of pronunciation that has not been submitted to empirical verification.
This study offers new empirical data in the assessment of the non-native speech intelligibility of a group of English learners of Spanish. Furthermore, we intend to analyse the possible impact of three individual variables - gender, aptitude and motivation - on our participants’ intelligibility scores in connected speech. While the study of individual differences within the area of Second Language Acquisition has proven fruitful (see our review of the literature below), the examination of those same individual variables with regard to the notion of non-native speech intelligibility is extremely scarce. Thus, this study will attempt to fill a research gap in the literature related to L2 phonological acquisition.
2 Review of the Literature
2.1 The Notion of Speech Intelligibility
It has proven difficult to offer a straightforward definition of speech intelligibility, given that the terminology is often used inconsistently across disciplines and authors. In the area of communication disorders, the term intelligibility is widely used (Rogers 1997), while in the field of L2 phonological acquisition, terms such as comprehensibility, communicative effectiveness, interpretability or accentedness are usually mentioned in connection with the notion of speech intelligibility. Weismer offers a thorough definition of this term when noting that intelligibility is a
relative measure of the degree to which a speaker’s speech signal is understood, the relativity depending at a minimum on the identities of the speaker and listener, what is spoken and where it is spoken. (Weismer (2008: 2)
Weismer’s definition gives an idea of the complexity of this notion, while underlining the wide array of factors that may have an impact on speech intelligibility.
In the area of L2 speech, Rogers (1997: 2-3) distinguishes between intelligibility, comprehensibility and communicative effectiveness. Intelligibility refers to the effective “production, transmission or perception by a listener of the speech sounds of a language”. Comprehensibility alludes to an individual’s success, or lack thereof, in conveying a specific message and includes not only variables related to the speech sounds but also to additional linguistic features. Communicative effectiveness, on the other hand, seems to encompass those linguistic and non-linguistic variables that may determine communicative success. Other authors, such as Gass and Varonis (1984), use the term comprehensibility to refer to communicative effectiveness without specifying the possible impact of linguistic and non-linguistic variables. Kenworthy equates intelligibility to understandability and defines this dimension as “being understood by a listener at a given time in a given situation” (Kenworthy 1996: 117). This broad proposal does not distinguish between those linguistic and non-linguistic variables that may have an influence on communicative success.
In the field of World Englishes, Smith and Nelson’s (1985) oft-cited proposal establishes a distinction between intelligibility, comprehensibility and interpretability. These dimensions refer to utterance recognition as well as understanding of locutionary and illocutionary force, respectively. Jenkins (2000) has also devoted a great deal of attention to the notion of intelligibility and the importance of phonological deviations for Interlanguage Talk (ILT). Intelligibility is seen by this author as a prerequisite for comprehensibility and interpretability. As Jenkins points out, intelligibility
concerns the recognition of the formal properties of words and utterances and, in particular, the ability to produce and receive phonological form but regards the latter as a prerequisite (though not a guarantee) of ILT success at the locutionary and illocutionary level. (Jenkins 2000: 78)
In the general field of L2 speech, Derwing & Munro distinguish between intelligibility, comprehensibility and accentedness. Intelligibility is defined as “the degree of a listener’s actual comprehension of an utterance”. Comprehensibility is operationalised in terms of “how difficult or easy it is to understand a given speech sample”, while accentedness is understood “as how different a pattern of speech sounds compared to the local variety” (Derwing & Munro 2009: 478-479). The definitions of these three dimensions clearly entail a departure from Smith and Nelson’s classical paradigm. Intelligibility includes both recognition of words and meaning, while the notion of speech comprehensibility is radically altered (Nelson 2008).
As seen from the sample of authors cited above, the definition of speech intelligibility has not remained consistent across disciplines. However, a general distinction can be established between the notion of recognition of linguistic form and meaning, and the communicative effectiveness of a speech act at the pragmatic level. Researchers generally focus on different aspects depending on their field of study and methodological decisions. The adoption of one perspective over another will obviously depend on the research goals, the type of instruments of data collection and the contextual factors surrounding a specific research project. In this study, intelligibility will be used to refer to the recognition and production of phonological form, similarly to Jenkins’s use of the term, but with the difference that production will point to our group of English learners of Spanish and recognition to a group of evaluators that consists only of native Spanish speakers
2.2 Measuring Speech Intelligibility: Methodological Considerations
It must be pointed out that intelligibility studies in the general field of communication disorders are more abundant than in the area of L2 speech. Munro (2008), in a recent literature review on foreign-accented speech, highlights the diversity of instruments used in measuring speech intelligibility, as well as the difficulty in eliciting appropriate speech samples. As far as elicitation techniques are concerned, researchers are presented with the dilemma of using controlled elicitation tasks (reading of words, sentences or passages) or using tasks aimed at obtaining more spontaneous speech samples (e.g. picture-description tasks). Both possess intrinsic advantages and disadvantages. Spontaneous or semi-spontaneous speech is more ecologically valid. However, it is also true that the elicitation of this type of speech samples makes any inter-participant comparisons extremely difficult. On the other hand, tasks consisting in the reading of lists of words, sentences or passages provide a higher degree of control over the lexical content. Furthermore, as noted by Munro,
if intelligibility is defined as the amount of a message that is actually understood, a comparison of the intended message with the received message is essential. (Munro 2008: 202)
This is extremely difficult to accomplish in the case of spontaneous or semi-spontaneous speech samples.
With regard to the actual measure of speech intelligibility, the majority of studies seem to opt either for rating scales or for orthographic transcriptions. It is noteworthy that the use of rating scales, a purely impressionistic way of assessment, makes it difficult to establish a distinction between intelligibility and perceived comprehensibility. If we opt for an orthographic transcription of speech samples, a decision regarding how to score evaluators’ transcriptions must also be made. Studies have varied between the transcription of words or sentences. The scoring procedures at the sentence level have ranged from the count of correctly transcribed words (e.g. Maasen & Povel 1985), correctly transcribed key or content words (e.g. Bradlow & Bent 2002), or correctly transcribed sentences as a whole (e.g. Benoît 1990). Rogers (1997), in her review of the literature, distinguishes between rating scales, phonetic transcription and identification tasks. This author concludes that, while rating scales do not offer information about the types of errors that may cause intelligibility problems, identifications tasks can be used as a diagnostic tool in explaining intelligibility deficits.
In summary, the chosen type of instrument and measurement procedure depends on the definition of speech intelligibility put forward by each researcher, as well as on the specific goals of each study. In our case, we centred on the assessment of intelligibility in connected speech. Speech samples were elicited through the use of the Spanish version of the Harvard Psychoacoustic Sentences. An orthographic transcription of the speech samples was used in the assessment of non-native speech intelligibility (see Section 3.3.3 below for a description of the scoring procedures).
2.3 The Role of Gender, Aptitude and Motivation in L2 Phonological Acquisition
One can say that success in SLA varies considerably from individual to individual. While all children achieve full mastery of their native language, except in severe cases of input deprivation, L2 learners experience variation in terms of rate of acquisition and ultimate attainment. In general, L2 learners reach lower levels of proficiency when compared to those exhibited by native speakers of the L2 (see Cook 2008 for criticisms on using a monolingual native speaker as the yardstick for success in Second Language Acquisition). The area of study that focuses on individual factors affecting the acquisition process has received a lot of attention in the past decades and has yielded numerous inconclusive and controversial results. The number of factors playing a role in Second Language Acquisition constitutes an ongoing debate. Different variables account for different amounts of variance depending on the researcher. This is not surprising considering that methodological shortcomings prevent, in most cases, a direct comparison between studies. In the area of pronunciation, Moyer (1999) cites formal instruction, L2 exposure, motivation, attitudes towards learning, personal success in linguistic and cultural assimilation, learning style, aptitude, attitude, and the use of input and feedback as factors affecting L2 phonological acquisition. Purcell and Suter (1980) stress the importance of aptitude for oral mimicry, length of stay in the target-language country or concern for pronunciation accuracy. More recently, Derwing (2008) mentions several factors including age, aptitude, motivation, phonological distance between L1 and L2, and degree of L2 exposure. We will now concentrate on the analysis of three specific individual variables: aptitude, motivation and gender.
In general, aptitude refers to one’s talent to acquire a new skill. Specifically, when talking about language aptitude, it alludes to an individual’s potential for learning a foreign language. It is, along with motivation, one of the most influential factors on language achievement (Ellis 2005). In the area of L2 phonological acquisition, the number of empirical studies focusing on the possible existence of a special ability in L2 pronunciation is rather scarce. As a result, Leather & James (as cited in Leather 1999: 11) mention that we do not have enough studies to support the idea that learners’ progress could be constrained by ‘biological differences in awareness of, and control over, the changing configurations of the articulators, or by differences in auditory sensitivities’. Additionally, Leather believes that the concept of a “good ear” for learning languages has not yet been incorporated as a variable in L2 speech (Leather 1999: 12). Some researchers have focused on the relationship between degree of foreign accent, “aptitude for oral mimicry” and musical ability. For example, authors such as Purcell and Suter (1980) or Flege et al. (1999) seem to have identified the ability of mimicking unfamiliar sounds as a significant predictor of the degree of foreign accent. Regarding the issue of musical ability, some researchers (Nakata 2002, Tanaka & Nakamura 2004) have reported a positive relation between pronunciation and musical ability. Slevc & Miyake, who claimed to have carried out “the first study that rigorously tested the musical-ability hypothesis and provided clear evidence for it” (Slevc & Miyake 2006: 679), assessed 50 native speakers of Japanese who arrived in the US after the age of 11, and concluded that musical ability is a significant predictor of L2 phonological ability (receptive and productive) even when controlling for other variables. However, it did not predict variance in syntax or lexical knowledge. One factor that has been widely studied in the past decades is the influence of working memory capacity on the acquisition of a second language. Initially, a distinction between working memory and short-term memory (STM) must be established. STM is the ability to store a small amount of information for a short period of time. Peterson & Peterson (1959: 194) estimate that, without rehearsal, STM fades quickly after about twelve seconds. Robinson defines working memory as “the workspace where skill development begins and where knowledge is encoded into and retrieved from long term memory” (Robinson 1995: 304). Working memory has been the object of some attention by Second Language Acquisition researchers in the past decade (Gass, Roots & Lee 2006, Van den Noort, Bosch & Hugdahl 2006). Some studies have specifically looked into the relationship between grammar and phonological STM (Ellis & Schmidt 1997). Others (Williams & Lovatt 2003) have demonstrated that phonological STM is connected to certain abstract aspects of grammar. Despite this surge in the number of studies, Gass & Selinker are surely right in pointing out that
working memory research in Second Language Acquisition is in its infancy. As with other constructs (e.g. competence), it is not always clear how best to measure it. (Gass & Selinker 2008: 272)
There is therefore a need for a better integration of working memory variables in the general context of Second Language Acquisition research.
Affective variables refer to a wide range of factors related to emotional and even cognitive aspects of an individual’s personality. Factors such as personality type, extraversion, foreign language anxiety, willingness to communicate, cognitive and learning style, to name a few, have yielded a great deal of empirical research within the general field of Second Language Acquisition (see Ortega 2009 for a review). Their study in connection with pronunciation accuracy or L2 phonological acquisition is much scarcer. In this section, we will look at one factor that most probably plays an important role in the phonological acquisition of a second language: motivation. Piske, MacKay & Flege (2001) concluded in their extensive literature review that motivation exerts some influence on degree of foreign accent. However, they also pointed out that
factors such as professional motivation, integrative motivation or strength of concern for L2 pronunciation accuracy does not automatically lead to accent-free L2 speech. (Piske, MacKay & Flege 2001: 12)
Furthermore, they did not consider motivational variables in their own experimental study by arguing that motivation has little effect when participants are immigrants in the L2 country. They do, however, concede that it may possess a stronger influence on those individuals who need to speak the target language without a foreign accent for professional reasons. From our perspective, the fact that motivational variables are difficult to define and to properly quantify does not necessarily mean that they do not possess a heavy influence on L2 phonological acquisition. In fact, Piller suggests that
age is not the critical factor in reaching high levels of L2 proficiency it is often assumed to be. Rather, personal motivation, choice and agency seem to be more crucial factors in ultimate attainment. Piller (2002: 23)
Piller’s analysis offers interesting insights when linking the concept of passing for a native speaker to actual performance and not to a “quality of being”. Oyama (1976) and Thompson (1991) are examples of studies that found no correlation between motivation and accuracy in pronunciation. On the other hand, there exist several studies that show a correlation between motivational variables and degrees of foreign accent. Suter (1976) tested, among other factors, the relation between motivation and performance in pronunciation. Sixty-one subjects from diverse linguistic backgrounds were required to fill out a questionnaire and undergo an interview. Participants had to mimic new sounds and produce free speech on a holiday experience. Additionally, they had to complete personality and aptitude tests. Results showed that the subjects’ concern towards pronunciation was one of the significant factors that could account for accuracy. Purcell & Suter (1980) carried out a more in-depth statistical analysis based on data from Suter’s study. Results revealed that concern for pronunciation accuracy accounted for 7% of the variance. Bongaerts et al. (1995, 1997) focused on highly advanced Dutch learners of English who had been exposed to extensive L2 input and received explicit training in English pronunciation. Results showed that, out of the 11 highly motivated participants, only 5 received scores similar to those obtained by native speakers. Similarly, Flege et al. (1995) did find that concern for pronunciation or integrative motivation were variables affecting the degree of perceived foreign accent. However, these factors only accounted for less than 3% of the variance. Moyer (1999) focused on highly motivated advanced English learners of German. After several tests that involved the reading of a list of words, sentences and a paragraph and being rated by four different judges, only one of the participants managed to pass as a native speaker. Professional motivation accounted for 41% of the variance. Moyer (2007) conducted an additional study with 50 immigrant learners of English in the US and concluded that accent is strongly linked to attitudes towards the target language such as reasons for learning English, perceived ability to improve in English or desire to improve one’s accent. In summary, results suggest that motivation does have an impact on L2 phonological acquisition. However, it is difficult to quantify the exact nature of such influence.
2.3.3 Gender Differences
Empirical research in L1 seems to suggest that there are certain gender differences with regard to pronunciation, i.e. women tend to exhibit better pronunciation and show a preference for more formal and prestige forms than their male counterparts (e.g. Silva Corvalan 2001). When we turn to the field of L2 phonological acquisition, research does not seem to confirm the existence of gender differences regarding pronunciation accuracy (e.g. Elliott 1995, Purcell & Suter 1980). It is also noteworthy that, in the past decade, some Second Language Acquisition researchers (e.g. Ohara 2001), following developments in areas such as sociolinguistics, have departed from the notion of gender as a biological construct, and have put forward a more dynamic conceptualisation based on the idea that gender is “something we do and not something we are” (Ehrlich 1997). On this issue, with regard to the potential impact of gender differences on pronunciation accuracy, Hansen Edwards notes the following:
When gender is framed and investigated as a biological construct, it does not seem to be a significant factor in L2 pronunciation accuracy. However, when gender is framed and investigated as a social construct, it does appear to impact the level of access learners have to L2 use opportunities and therefore the ability to get L2 input and negotiate meaning, which appear to affect L2 development. Hansen Edwards (2008: 255)
It would seem, therefore, that when gender is conceptualised as a social construct, biologically-based factors interact with other variables that may arise as a result of different contextual situations. Given the limitations, constraints and research goals of this study, the notion of gender has been conceptualised here as a mere biological distinction between males and females. It is clear that the choice between a biological construct and a more socially-dynamic notion of gender is determined by the type of study, the contextual factors surrounding the research and the different methodological choices.
Twenty secondary-school students (Year 11) from a mixed-ability Spanish language class participated in this study. Participants belonged to an average size secondary school in Hertfordshire, England. Only English learners of Spanish with no hearing or speech impediments, no prolonged periods of stay in Spanish speaking countries and no exposure to Spanish at home were considered for this study. Twenty-two participants met initially our set of criteria. Two of them did not complete all the recording sessions and were therefore excluded from this study.
Sixty evaluators (33 male and 27 female), native speakers of Peninsular Spanish, were recruited to assess the speech samples collected in this study. Listener-related variables, such as familiarity with foreign accents, previous training or linguistic background, have proven to have an effect on intelligibility and comprehensibility ratings (Derwing & Munro 2009, Kennedy & Trofimovich 2008, Rogers 1997). In order to neutralise the effect of some of these factors, we decided to recruit naïve listeners with no hearing or speech impediments, from different social and educational backgrounds, a broad age range, with no regular contact with English speakers and with no extended periods of stay in English-speaking countries.
All subjects were native Spanish speakers with no reported hearing or speech impediments. Their mean age was 26.5. In terms of educational background, 35% held university degrees, 15% were studying towards their Bachelor's Degree, 30% held secondary school qualifications (Bachillerato, Formación Profesional) and 20% had either diverse continuing education certificates or had completed the E.G.B. (Enseñanza General Básica). None of them had regular contact with English speakers and they did not report on any prolonged stays in English speaking countries. On occasions, they did report on holiday periods spent in English speaking countries but always for a maximum of up to two weeks. As far as knowledge of foreign languages was concerned, 45% of evaluators did not speak any foreign languages at all, 30% claimed to speak French at a beginner or intermediate level of proficiency, while 25% claimed to possess a beginner level of proficiency in English. It is worth pointing out that any individuals who reported an intermediate or advance level of proficiency in English were automatically excluded from this study.
3.3.1 Assessing Motivation
Students completed a motivation questionnaire (Appendix 1) to assess their level of motivation towards the learning of Spanish pronunciation. The motivation questionnaire was adapted from Wen (2005), which, in turn, was based on Scherer (1984) and Gardner's Attitude / Motivation Test Battery (AMTB). The test consisted of 18 questions, which were rated on a scale of 1 to 10 (‘strongly agree to strongly disagree’). Wen’s questionnaire was originally designed to study learners’ motivation in a Second Language environment, i.e. an environment in which students are learning the L2 in the target-language country. Thus, only questions applicable to contexts related to foreign language learning were used in this study. Questions 3, 8, 12, 13, 16, 18, 19, 20, 22, 23, 24, 25, 27 and 28 were eliminated altogether from the original questionnaire and questions 11, 14, 17 and 21 were adapted to make them more suitable for the specific population under study. Nevertheless, the same four areas of motivation were examined: novelty, i.e.
- students’ reaction to new material,
- pleasantness, which alludes to the pleasantness or unpleasantness that arises out of a stimulus,
- coping material, which refers to a student’s ability to adjust to a specific stimulus, and
- goal significance, i.e. the relevance of a specific stimulus with regards to an individual’s goals (see Schuman 1999 for a detailed analysis of models of stimulus appraisals).
It is also noteworthy that we reduced the Liker scales from 10-point to 5-points.
This questionnaire can be criticised for a number of reasons. It is true that in the past 15 years, the study of motivation has veered towards dynamic and more qualitative analysis of this construct (see Ortega 2009 for a review). However, given the number of participants and the multiple constraints that we encountered when carrying out this research project, it was considered that a quantitative questionnaire of this kind was the most effective way to obtain a measurement of students’ level of motivation.
3.3.2 Measuring Aptitude
220.127.116.11 Measuring Working Memory Capacity
In this study, aptitude will be operationalised as a measurement of students’ working memory capacity and their ability for oral mimicry. As an instrument for measuring students’ working memory, we chose a battery of tests designed by Lewandowsky et al. (2010) for MATLAB (Matrix Laboratory). Tests included a sentence-span task, an operation-span task a spatial short-term memory task and a memory updating task. As the authors point out:
These tasks were chosen in order to provide a heterogeneous set of measures of working memory capacity, thus reducing method variance and tapping into two content domains of working memory (verbal, including numerical, vs. spatial) and two of its functional aspects (storage in the context of processing and relational integration. (Lewandowsky et al.: 2010: 571)
Moreover, these tasks offer a measurement of both students’ memory capacity and processing ability. A full description of each task plus detailed scoring procedures can be found in Lewandowsky et al. (2010):
The memory updating task presents participants with a series of digits that students attempt to remember while performing a series of arithmetic operations. A total of 15 series of digits are presented. Sequences are generated at random.
The operation span task presents participants with a sequence of arithmetic equations (e.g. 10- 4 = 6) and a number of consonants for subsequent recollection. First, candidates need to indicate if the equation is correct or incorrect. After each equation, a consonant appears on the computer screen for 1 sec. The goal is not only to judge if the equation is right or wrong, but also to recall all consonants presented at the end of the series. All sequences are generated at random.
The sentence span task is similar to the previously described operation span task. In this case, the processing element of the test centres on determining whether a series of phrases are meaningful or meaningless (see Lewandosky et al. 2010 for a full description of the type of sentences included as stimuli).
In the spatial short-term memory task, participants are presented with a series of dots in a matrix on the computer screen. Dots appear one by one within different cells of the matrix for 900 milliseconds. Participants need to recall their patterns of presentation, i.e. the absolute position of each dot is irrelevant and what becomes important is the recollection of their overall patterns of display.
With regard to the scoring procedures, the indications provided by Lewandowsky et al. (2010) were followed at all times. Nevertheless, it is worth pointing out that all results were transformed into percentages in an attempt to facilitate comparisons between tasks and groups of students.
18.104.22.168 Measuring Ability for Oral Mimicry
We followed the procedures set by Lord (2006) in the design of an oral mimicry task. The main researcher in this study - a native speaker of Spanish - read out loud ten sentences (Appendix 2). Each sentence contained one invented word that was phonotactically possible in Spanish. Words acted as nouns or adjectives in the carrier sentences and ranged from 2 to 4 syllables in length. Students were recorded, repeating each sentence immediately after the researcher. Students recordings were subsequently transcribed, and a score was calculated based on the number of correctly produced invented words (out of a maximum of 10).
3.3.3 Assessing Intelligibility
A transcription task involving speech samples elicited through the reading of the Spanish version of the Harvard Psychoacoustic Sentences was used to assess intelligibility in connected speech. The Harvard Psychoacoustic Sentences (Egan 1948) are widely used to assess the intelligibility of words in sentence context. They consist of a group of phonetically-balanced sentences with varied and correct syntactic structures. Evaluators do not choose between several options but, rather, simply transcribe or repeat the sentence they have heard. This task is easy to administer and score. However, it does present some important limitations, as noted by Gibbon et al. (1997), such as a restricted number of items, a strong learning effect or a high percentage of correct answers. This is due to the fact that the evaluators are able to use not only phonetic but also semantic information during the transcription task.
The Spanish version of this test (Valero 1991) consists of ten groups of ten sentences each (Appendix 3). Only the first five series will be used to assess intelligibility at this level of analysis. Sayings, colloquial and idiomatic expressions as well as interrogative or exclamation sentences have been excluded from the corpus. Sentences are syntactically well-formed and include simple and complex sentences, both coordinated and subordinated, as well as all verb tenses and grammatical persons. Each group of ten sentences is phonetically balanced. With regard to phoneme frequency of occurrence, data from Navarro Tomás (1946) and Alarcos (1965) has been used when compiling the list of stimuli. It must also be pointed out that Navarro Tomás does not offer any details on frequency of occurrence of allophones and, thus, the balance of the corpus must be considered in a phonological and not in a phonetic sense.
As far as the assessment of the Harvard Psychoacoustic Sentences is concerned, two different scoring criteria could a priori be adopted: (i) a strict criterion that would penalise those errors that can be explained through contextual information and (ii) a more lenient criterion that would disregard this type of errors. For example, if one of the evaluators transcribes “los libro están sobre la mesa” and “libros” is a key word, adopting a strict scoring criterion would consider this instance as an error. Additionally, there is also the option of taking into consideration only those errors committed on key words or those committed on any word. Therefore, the following alternatives were initially examined:
- To consider only errors in the production of key words without taking into consideration errors that could be explained from the context (as explained above).
- To consider errors in key words and to penalise any mistake.
- To consider errors in any word without taking into consideration the mistakes that could be explained from the context.
- To consider those errors in any word and to penalise all mistakes.
After examining the test results, we decided to follow the first of the aforementioned criterion. Due to the nature of the stimuli involved (meaningful sentences with correct syntactic structures), it is unlikely that evaluators would write sentences without respecting basic grammatical rules. Evaluators, when transcribing the answers, look for meaningful grammatical patterns and, thus, a strict corrective criterion would offer no substantial variations on intelligibility scores. Results are therefore presented by determining the number of key words that have been correctly transcribed for each sentence. Key words are those with a clear lexical content. Modal or auxiliary verbs, adverbs, pronouns, prepositions, conjunctions, articles, and determiners have been excluded. All scores were transformed into percentages to facilitate inter-task comparisons.
4.1 Recording Sessions
All recording sessions of students’ oral productions were made on a pc / laptop equipped with an M-Audio MobilePre USB interface. Each session ran individually in a quiet room in the school premises. The microphone was placed 30 cm from each participant’s mouth. Subjects were instructed not to speak directly into the microphone. We used ProRec 1.2 (developed by Mark Huckavale University College London and freely available on the UCL website) as audio recording software. Stimuli were presented to students on the computer screen and were recorded directly to disk from the microphone, using a 48 KHz. sampling rate. Files were saved in a WAV format. A translation of stimuli into English was also provided to students in a paper format. Recording sessions were carried out as follows:
- Preliminary explanation of the task
- Questions and answers
- Testing of the recording device
- Recording of practice stimuli
- Clarification of possible questions
- Recording of real stimuli on the computer
It is important to point out that enough time was given to participants to read the selected sentences in its entirety and to consult the English translation if necessary.
4.2 Listening Sessions
Each participant in this study was randomly assigned to a panel of three evaluators (a total of 15 panels). The presentation of stimuli was randomised to avoid any possible learning effects. Stimuli were presented through headphones, and all transcription tasks were completed on the computer, using the software interface provided by E-Prime 2.0. Each evaluation session was carried out as follows:
- Preliminary reading of the instructions on how to perform the task;
- Oral explanation of the task by the researcher and answers given to any possible questions;
- Clarification of any possible questions;
- Presentation of the real stimuli.
The presentation of stimuli consisted of 50 sentences from the Spanish version of the Harvard Psychoacoustic Sentences. Evaluators had to orthographically transcribe each sentence.
5 Analysis of Results and Discussion
Intelligibility scores in connected speech ranged from 64.70% to 96.29%, with a standard deviation of 7.87 and a mean of 85.62. The degree of intelligibility loss for the population of L2 learners under study therefore reached an average of 14.38%. This certainly warrants further studies focusing on determining the degree of speech intelligibility of different populations of L2 learners.
Table 1 below presents the participants’ results of our motivation questionnaire. The questionnaire consisted of 18 statements that students rated on a scale of 1 to 5 (‘strongly disagree to strongly agree’). Results were calculated by averaging the score yielded by the Likert scale for each statement. They were later converted into percentages to facilitate comparisons, a higher percentage indicating a higher degree of motivation. Descriptive statistics showed that scores ranged from 48 to 90 with a mean of 68.60, a standard deviation of 12.88. Skewness and kurtosis values were 0.15 and 1.13, ensuring the normality of the distribution. No statistically significant differences in terms of motivation were recorded between male and females (t = -0.53, p = 0.60).
Aptitude was measured through a combination of an oral mimicry test and a battery of tasks to measure students’ working memory capacity. As pointed out above, there were ten tokens in the oral mimicry section. Scores ranged from 40 to 90, with a mean of 66.25 and a standard deviation of 17.23. Table 2 below presents students’ scores in the battery of tests that served to measure working memory capacity. Results were quite similar for the updating memory, sentence span and operation-span tasks. Results also showed that scores on the spatial short-term memory task were significantly higher than those in the other three subtests. Overall working memory capacity was calculated as an average of participants’ scores in all four subtests. Descriptive statistics for the overall score are as follows: scores ranged from 46.75 to 87.13 with a mean of 67.09 and a standard deviation of 13.66. The values for skewness and kurtosis are -0.02 and –1.28, respectively. Table 2 also presents an overall score for language aptitude, an average of the oral mimicry task and the battery of working memory tasks.
It is important to note that an examination of possible differences in intelligibility results according to gender revealed that no statistically significant differences were found with regard to intelligibility scores in connected speech (t (18) = -0.66, p = 0.51).
Table 3 below presents the correlation between aptitude, motivation and intelligibility scores in connected speech. Pearson’s r was used to perform the different correlation analyses. In this study, significant but moderate correlations were found between motivation and intelligibility scores in connected speech (r = 0.45). Correlation between aptitude and degree of speech intelligibility was slightly higher (r = 0.63).
One of the goals of this study is the exploration of a possible correlation between certain individual differences (motivation, aptitude and gender) and participants’ intelligibility scores in connected speech. Gender, operationalised from a biologically point of view, did not prove to be correlated with intelligibility scores. In other words, no statistically significant differences between males and females were detected in intelligibility scores. Results confirm previous studies exploring the relationship between pronunciation accuracy and gender (e.g. Elliot 1995).
With respect to aptitude and motivation, it is important to note that, in the general field of Second Language Acquisition, both variables account for a significant degree of variance in learners’ achievement. As noted by Dörnyei and Skehan (2003), correlations usually range between 0.20 and 0.60. Specifically, as far as the acquisition of L2 phonology is concerned, certain studies have linked the degree of motivation to pronunciation accuracy or perceived degree of foreign accent (e.g. Moyer 2007), while others have only been able to show a very low correlation between both variables (Flege et al. 1995) or even no correlation whatsoever (Thompson 1991). In the case of aptitude, both aptitude for oral mimicry and working memory capacity or phonological working memory have been linked to pronunciation accuracy (e.g. Flege et al. 1999; Purcell and Suter 1980). In this study, motivation was operationalised as students’ scores on a motivation questionnaire, while aptitude was quantified as a combination of scores from a battery of tests that served to measure our participants’ working memory capacity, as well as from a test that was designed to measure their ability for oral mimicry. Both aptitude and motivation were significantly correlated with intelligibility scores. In general, the correlation coefficient was higher than in previous studies, in which the variables involved were pronunciation accuracy or degree of perceived foreign accent. As far as aptitude is concerned, it is not surprising to find a significant correlation between degree of intelligibility and ability for oral mimicry, as well as working memory capacity. It is, in fact, quite consistent with those studies that have explored the impact of aptitude and L2 phonological acquisition (e.g. Flege et al. 1999, Purcell & Suter 1980). With regard to motivation, the level of correlation yielded by this study is consistent with the coefficients found in other areas of Second Language Acquisition (see Ellis 2005).
As pointed out in our review of the literature, research on individual variables reveals that L2 phonological acquisition is not solely a linguistic matter, but there are a number of extralinguistic aspects that play a major role during the learning process. Our empirical results confirm that the degree of speech intelligibility benefits from a high level of motivation and a high degree of aptitude, while gender does not seem to exert any influence on intelligibility scores. In addition, the amount of L2 use, the amount of L2 input and the age at the onset of the acquisition process are also variables that have proven to be important in the process of L2 phonological acquisition. In this study, aptitude was operationalised as a combination of ability for oral mimicry and working memory capacity. It is noteworthy that there are very few proposals for the treatment of working memory capacity in the foreign language classroom. It seems logical to assume that working memory capacity may play an important role in communicative teaching contexts in which students must process a large amount of oral information. The working memory may also play an important role in processing written and audiovisual material. In addition, it has also been posited that the relationship between L2 competence and working memory capacity is based on the idea that an increased capacity of the latter will allow L2 learners to release a series of cognitive resources that would otherwise be engaged in input processing. On this basis, one can assume that exercises that increase overall L2 fluency could compensate for a low working memory capacity. Priority should be given to classroom activities that will help improve the automatic recognition of words. Moreover, we believe it extremely useful to offer learners enough opportunities to process, reproduce and automate the production of words and structures. Additionally, we could also make use of activities to improve strategies aimed at linking pre-existing L1 and L2 knowledge to new L2 input. If we increase the use of what we might call long-term knowledge, the burden on working memory will be drastically reduced. Other techniques that may serve to compensate for a low working memory capacity include the use of visual aids. Visual input can be used to replace the working memory’s temporary storage function allowing for the processing of a greater amount of information. It would be advisable for teachers to increase their learners’ use of written and visual aids during their communicative interactions, especially in the early stages of the learning process. Our knowledge of the role of the working memory in foreign language learning is still relatively small. Further research is certainly needed to determine whether teaching techniques that focus on an increasing working memory capacity can have a positive impact on L2 learning. We also need more research to assess whether educational intervention techniques aimed at compensating for a reduced working memory capacity can actually improve L2 learning. The use of exercises designed to improve students’ degree of fluency or the use of visual and written aids are just some of the pedagogical tools that can lead to an improvement in L2 phonological acquisition. All these suggestions will need empirical validation and should offer possible avenues for further research in the field.
With regard to motivation, L2 learners should be encouraged to develop positive attitudes towards pronunciation. It is important that they become aware of the positive effects that an improved pronunciation may have on communication. The better instructors are aware of these variables, the better they will be able to anticipate the level of proficiency their students may reach. In other words, they will be able to predict their students’ progress in L2 phonological acquisition with a reasonable degree of reliability.
It is noteworthy that the amount of research devoted to intelligibility studies within the area of Second Language Acquisition is rather scarce. Further research should be able to provide a more thorough picture of the variables involved in intelligibility loss. We also need further intelligibility studies that focus on different language combinations. Moreover, special attention should be devoted to the relation between intelligibility at the word level and in connected speech. In particular, it seems essential to gain a better understanding of the multiple factors that have an impact on intelligibility in spontaneous conversation.
Appendix 1: Motivation Questionnaire (adapted from Wen 2005)
Read each statement and circle the number that best reflects your opinion
1= Strongly disagree
5= Strongly agree
1. Spanish is not important for me to learn about because it is not necessary in the world.
2. I feel happy if people tell me that I have great pronunciation.
3. The majority of my experience of learning Spanish has been pleasant.
4. I like to watch TV, listen to the radio, listen to the songs or watch movies in Spanish.
5. I want to pronounce as best as I can.
6. Pronouncing like a NS of Spanish is important for me.
7. I seldom watch or listen to any kind of Spanish programmes.
8. Better pronunciation helps me in my career plan or study.
9. I seek chances to speak Spanish.
10. I pay careful attention to how Spanish speakers pronounce words.
11. When somebody teases me for my pronunciation, I feel a little embarrassed but it’s ok.
12. I avoid speaking Spanish if somebody teases me for my pronunciation.
13. If somebody teases me for my pronunciation, it won’t discourage me from learning Spanish; on the contrary, I will do my best to improve in order to change their opinions.
14. I like to imitate how Spanish speakers pronounce Spanish on TV, radio or movies.
15. I would like to have Spanish speaking friends.
16. I don’t like to speak Spanish.
17. I don’t care about my pronunciation.
18. I like to speak Spanish.
Appendix 2: Oral Mimicry Task (adapted from Lord 2006)
Sentences (invented words in italics)
1- Se podía oír a través de la ventana el violento pristonar contra los cristales.
2- El pequeño fatusal desapareció rápidamente detrás de los árboles
3- Me gustaría ver el blasofón que tienes en la cocina después de la cena
4- Para ser un buen blugón, se necesita estudiar mucho
5- El crestalar hacía muy difícil que pudiéramos concentrarnos
6- Me dijo que el sobredino estaba a punto de llegar
7- Tuvo que devolver a la policía el plautón que encontró cerca de mi casa
8- Me dijo que tuviera cuidado con el blaito ya que era peligroso
9- El clarucio era lo único que los mantenía vivos
10- Abre el estubre que está sobre la mesa
Appendix 3: Spanish Version of the Harvard Psychoacoustic Sentences (Egan 1948; Valero 1991)
La oscuridad me da mucho miedo.
Podéis venir con nosostros al cine.
Se ha roto la tapa del bote de mermelada.
Ganará lo que ganaba en su anterior trabajo.
Carlos es capaz de no saludarnos.
Se tiene por forzuda, pero no resiste nada.
Las llaves están en mi bolsa.
Pili quiere que tú y yo cojamos zarzamoras.
Le apatece un poco de vino blanco.
Las notas de diseño que te faltan no las sabrás por mí.
He comido carne con patatas fritas.
El balazo se desvió de la diana.
Se pegaron en medio de la curva.
Esta sustancia nos prolonga la memoria.
Recoge su llavero y pónselo dentro del cesto.
Las oscuras pestañas enmarcan su dulce mirar.
Una de las redes ya está en mi barco pesquero.
Te vencí y di la copa al orfelinato.
Beber anís puede quemarte el gaznate.
Pon mis discos y cállate.
Las hormigas se han comido las rosas.
Le envié un dinero por carta certificada.
Las vallas de metal destrozan el puente.
Yo quiero un poco más de jabón.
El duro capataz nos riñó bastante.
Carlos no ha pagado el mes de enero.
Lucía nos critica siempre a David.
Los techos blancos suelen necesitar pintura.
Me enfadé porque llegó mal vestido.
Lázaro casi anda más pronto que gatea.
Cuando te vaya bien, llámame.
Los roperos son pequeños.
La tila te calma y duermes mejor.
Si no cenas tendrás dolor de barriga.
La pieza que no funciona es del carburador.
No tienes que soportarle sus caprichos.
Las tazas de café no están dónde siempre.
Olga no tiene valor para quedarse.
Si la luz te incomoda, te vas a dormir.
Su nieta Lola también sabe polaco.
Puedes fumar, pero vete al balcón.
La tía Carmen quería guisar el conejo.
No des patadas a los rosales.
Necesito un centímetro para medir las dos telas.
Con un quilo de boniatos ya tengo de sobra.
Quizás pueda volver a llamar mañana por la noche.
Si quieres rezar, te dejo solo.
Nunca debí dinero ni lo tomé prestado.
Sus cacerías acabaron con la fauna de la zona.
Me diste la paga de este mes.
Alarcos, E. (1965) Fonología española. Madrid: Gredos.
Benoît, C. (1990) An intelligibility test using semantically unpredictable sentences: towards the quantification of linguistic complexity. Speech Communication 9, 4: 293-304.
Bent, T. & Bradlow, A. R. (2003) The interlanguage speech intelligibility benefit. Journal of the Acoustical Society of America 114, 3: 1600-1610.
Bisquerra, R. (ed.) (2004) Metodología de la investigación educativa. Madrid: La Muralla.
Bongaerts, T., Planken, B. & Schils, E. (1995) Can late learners attain a native accent in a foreign language? A test of the Critical Period Hypothesis. In D. Singleton and Z. Lengyel (eds.) The Age Factor in Second Language Acquisition. Clevedon: Multilingual Matters. pp. 30-50.
Bongaerts, T., Summeren, C., Planken, B. & Schils, E. (1997) Age and ultimate attainment in the production of foreign language. Studies in Second Language Acquisition 19: 447-465.
Cook, V. (2008) Second Language Learning and Language Teaching. London: Arnold.
Derwing, T. M. (2008) Curriculum issues in teaching pronunciation to second language learners. In J. G. Hansen Edwards and M. L. Zampini (eds.) Phonology and second language acquisition. Amsterdam: John Benjamins. pp. 347-369.
Derwing, T. M. (2008) Curriculum issues in teaching pronunciation to second language learners. In J. G. Hansen Edwards and M. L. Zampini (eds.) Phonology and second language acquisition. Amsterdam: John Benjamins. pp. 347-369.
Derwing, T. M. & Munro, M. J. (2009) Putting accent in its place: Rethinking obstacles to communication. Language Teaching 42: 276-490.
Dörnyei, Z. & Skehan, P. (2003) Individual differences in second language learning. In C. J. Doughty and M. H. Long (eds.) The handbook of second language acquisition. Oxford: Blackwell. pp. 589-630.
Egan, J. P. (1948) Articulation testing methods. Laryngoscope 58, 9: 955-991.
Ehrlich, S. (1997) Gender as social practice: Implications for second language acquisition. Studies in Second Language Acquisition 19: 421-446.
Elliott, A. (1995) Field independence/dependence, hemispheric specialization, and attitude in relation to pronunciation accuracy in Spanish as a foreign language. Modern Language Journal 79: 356-371.
Ellis, R. (2005) Instructed Second Language Acquisition: A literature review. Wellington: Ministry of Education New Zealand.
Ellis, N. & Schmidt, R. (1997) Morphology and longer distance dependencies: laboratory research illuminating the A in SLA. Studies in Second Language Acquisition 19: 145-171.
Flege, J. E., Munro, M. J. & MacKay, I. (1995) Factors affecting degree of perceived foreign accent in a second language, Journal of the Acoustical Society of America 97: 3125-3134.
Flege, J. E., Yeni-Komshian, G. H. & Liu, Serena (1999) Age constraints on second-language acquisition. Journal of Memory and Language 4: 78-104.
Gass, S. & Selinker, L. (2008) Second Language Acquisition: An Introductory Course. New York,NY: Routledge.
Gass, S. & Varonis, E. (1984) The effect of familiarity on the comprehensibility of nonnative speech. Language Learning 34: 65-89.
Gass, S., Roots, R. & Lee, J. (2006) Inhibition and working memory capacity in a second language. Paper presented at the 16th European Second Language Association Conference, Antalya, Turkey.
Hansen Edwards, J. G. (2008) Social factors and variation in production in L2 phonology. In J. G. Hansen Edwards and M. L. Zampini (eds.) Phonology and Second Language Acquisition. Amsterdam: John Benjamins Publishing Company. pp 251-279.
Jenkins, J. (2000) The phonology of English as an international language: new models, new norms, new goals. Oxford: Oxford University Press.
Kenworthy, J. (1996) Teaching English Pronunciation. Harlow: Longman.
Leather, J. (1999) Second Language Speech Research: An Introduction. Language Learning 49, S1: 1-56.
Leather, J. & James, A. R. (1991) The acquisition of second language speech. Studies in Second Language Acquisition 13: 305-341.
Levis, J. (2005) Changing contexts and shifting paradigms in pronunciation teaching. TESOL Quarterly 39, 3: 369-377.
Lewandowsky, S., Oberauer, K., Yang, L.-X., and Ecker, U. K. H. (2010) A working memory test battery for Matlab. Behavioral Research Methods 42: 571-585.
Lord, G. (2006) Defining the indefinable: Study abroad and phonological memory abilities. In C. A. Klee and T. L. Face (eds.), Selected proceedings of the 7th conference on the acquisition of Spanish and Portuguese as first and second languages. Somerville, MA: Cascadilla Proceedings Project. pp. 40‐46.
Maasen, B., & Povel, D.-J. (1985). The effect of segmental and suprasegmental corrections on the intelligibility of deaf speech. Journal of the Acoustical Society of America, 78, 877-886.
Meyerhoff, M. (2011) Introducing Sociolinguistics. New York: Routledge.
Moyer, A. (1999) Ultimate attainment in L2 phonology: The critical factors of age, motivation and instruction. Studies in Second Language Acquisition 21: 81-108.
Moyer, A. (2007) Empirical Considerations on the Age Factor in L2 Phonology. Issues in Applied Linguistics 15, 2: 109-127
Munro, M. J. (2008) Foreign accent and speech intelligibility. In J. G. Hansen Edwards and M. L. Zampini (eds.) Phonology and Second Language Acquisition. Amsterdam: John Benjamins. pp. 193-218.
Munro, M. J. & Derwing, T. M. (2011) The foundations of accent and intelligibility in pronunciation research. Language Teaching 44: 316-327.
Nakata, H. (2002) Correlations between musical and Japanese phonetic aptitudes by native speakers of English. Reading Working Papers in Linguistics 6: 1-23.
Navarro Tomás, T. (1946) Escala de frecuencia de los fonemas españoles. Estudios de fonología española. Syracuse, 15–30.
Nelson, C. L. (2008) Intelligibility since 1969. World Englishes 27: 297-308.
Ohara, Y. (2001) Finding one’s voice in Japanese: A study of the pitch levels of L2 users. In A. Pavlenko, A. Blackledge, I. Piller, and M. Teutsch-Dwyer (eds.) Multilingualism, second language learning, and gender. Berlin: Mouton de Gruyter. pp. 231-256.
Ortega, L. (2009) Understanding second language acquisition. London: Hodder Arnold.
Oyama, S. (1976) A sensitive period for the acquisition of a non-native phonological system. Journal of Psycholinguistic Research 5: 261-283.
Peterson, L. R. and Peterson, M. J. (1959) Short-term retention of individual verbal items. Journal of Experimental Psychology 58: 193-198.
Piske, T., MacKay, I. R. A. & Flege, J. E. (2001) Factors affecting the degree of foreign accent in an L2: A review. Journal of Phonetics 29, 2: 191-215.
Purcell, E. & Suter, R. (1980) Predictors of pronunciation accuracy: A reexamination. Language Learning 30, 2: 271-87.
Robinson, P. (1995) Attention, memory, and the “noticing” hypothesis. Language Learning 45: 283-331.
Rogers, C. L. (1997) Intelligibility of Chinese-accented English. Unpublished doctoral dissertation. Indiana University.
Silva Corvalán, C. (2001) Sociolingüística y pragmática del español. Washington, DC: Georgetown University Press.
Slevc, L. R., & Miyake, A. (2006) Individual differences in second language proficiency: Does music ability matter? Psychological Science 17: 675-681.
Smith, L. E. & Nelson, C. (1985) International intelligibility of English: Directions and resources. World Englishes 4: 333-342.
Suter, R. (1976) Predicators of Pronunciation accuracy in second language learning. Language Learning 26: 233-53.
Tanaka, A. & Nakamura, K. (2004) Auditory memory and proficiency of second language speaking: A latent variable analysis approach. Psychological Reports 95: 723-734.
Thompson, I. (1991) Foreign Accents Revisited: The English Pronunciation of Russian Immigrants. Language Learning 41: 177-204.
Valero, A. (1991) El corpus de las frases psicoacústicas de Harvard: una adaptación al castellano. Unpublished manuscript. Universitat Autònoma de Barcelona, Departamento de Filología Española.
van den Noort, M., Bosch, P. and Hugdahl, K. (2006) Foreign language proficiency and working memory capacity. European Psychologist 11: 289-296.
Weismer, G. (2008) Speech intelligibility. In M. J. Ball, M. Perkins, N. Müller and S. Howard (eds.) The Handbook of Clinical Linguistics. Oxford: Blackwell Publishers. pp. 568-582.
Wen, T. (2005) The Role of Motivation in Second Language Pronunciation. Unpublished master’s thesis. University of North Texas. Williams, J. N. and Lovatt, P. J. (2003) Phonological Memory and Rule Learning. Language Learning 53, 1: 67-121.
Ángel Osle Ezquerra
Queen Mary, University of London
Department of Iberian and Latin American Studies
Mile End Rd., London, E1 4NS
 This battery of tests is freely available for download on www.cogsciwa.com. It is noteworthy that, in order for the programme to work, a version of MATLAB must be previously installed on each computer, along with the Psychophysics Toolboox.
 The Pearson correlation coefficient being designed for quantitative variables, is an index that measures the degree of co-variation between two variables. Pearson’s r is an index of straightforward calculation and interpretation. Its absolute values range between –1 and +1. The magnitude of the relationship is specified by the numeric value of the coefficient, the sign reflecting the direction of such value. There are, however, no precise guidelines as to how to interpret this correlation coefficient. Much depends on the nature of the investigation and the size of the sample. In general terms, Bisquerra (2004) indicates that a correlation between 0.4 and 0.6 can be considered as moderate, while between 0.6 and 0.8 as high.