ORIGINAL RESEARCH published: 10 June 2021 doi: 10.3389/fcomm.2021.660917 When the Easy Becomes Difficult: Factors Affecting the Acquisition of the English /iː/-/ɪ/ Contrast Juli Cebrian 1*, Celia Gorba 1 and Núria Gavaldà 2 1 Departament de Filologia Anglesa i Germanística, Facultat de Filosofia i Lletres, Universitat Autònoma de Barcelona, Barcelona, Spain, 2Facultad de Educación, Universidad Internacional de la Rioja, Logroño, Spain The degree of similarity between the sounds of a speaker’s first and second language (L1 and L2) is believed to determine the likelihood of accurate perception and production of the L2 sounds. This paper explores the relationship between cross-linguistic similarity and the perception and production of a subset of English vowels, including the highly productive /iː/-/ɪ/ contrast (as in “beat” vs. “bit”), by a group of Spanish/Catalan native speakers learning English as an L2. The learners’ ability to identify, discriminate and produce the English vowels accurately was contrasted with their cross-linguistic perceived similarity judgements. The results showed that L2 perception and production accuracy was not always predicted from patterns of cross-language similarity, particularly regarding the difficulty distinguishing /iː/ and /ɪ/. Possible explanations may involve the way the L2 /iː/ and /ɪ/ categories interact, the effect of non-native acoustic cue reliance, and the roles of Edited by: John Archibald, orthography and language instruction. University of Victoria, Canada Keywords: vowel contrast, L2 perception, L2 production, cross-linguistic similarity, individual variation Reviewed by: Philip J. Monahan, University of Toronto, Canada Anabela Rato, INTRODUCTION University of Toronto, Canada Learning a second or foreign language (L2) after childhood is a difficult task for a variety of reasons. In *Correspondence: Juli Cebrian addition to learner-related factors such as age of learning and amount of first and second language use juli.cebrian@uab.cat (Piske et al., 2001; Flege et al., 2003, among others), the existence of a first language (L1) phonological system in part accounts for the fact that adult L2 speakers rarely sound like native speakers Specialty section: (Trubetzkoy, 1939; Rochet, 1995; Strange, 1995, among others). Variability in L2 performance has This article was submitted to been linked to the degree of similarity between first and second language sound systems (Lado, 1957; Language Sciences, Best, 1995; Flege, 1995; Kuhl and Iverson, 1995, among others). Models of L2 speech relate the a section of the journal likelihood of accurate perception and production of target language phones to the degree of similarity Frontiers in Communication between L1 and L2 phones. Early models of L2 speech, such as the contrastive analysis hypothesis Received: 29 January 2021 (Lado, 1957), suggested that the closer a target language phone was to a L1 category, the easier it was to Accepted: 04 May 2021 perceive and produce it accurately. Escudero’s (2005) Second Language Linguistic Perception Model Published: 10 June 2021 (L2LP) posits that L2 phones that have a similar L1 counterpart are easier to learn than new phones Citation: with no clear counterpart in the L1. The L2LP claims that the L2 is, originally, a copy of the L1, and this Cebrian J, Gorba C and Gavaldà N copy may then evolve toward more target-like values. Thus, learning a similar sound involves an (2021) When the Easy Becomes adjustment of the category boundaries and their acoustic properties, whereas acquiring a new sound Difficult: Factors Affecting the Acquisition of the English /iː/-/ɪ / requires establishing a new L2 to L1 mapping prior to the process of phonetic adjustment. Other recent Contrast. theories link successful L2 category creation to the ability to distinguish between L1 and L2 sound Front. Commun. 6:660917. categories. Best and colleagues’ Perceptual Assimilation Model (PAM, Best, 1995), and its adaptation to doi: 10.3389/fcomm.2021.660917 L2 speech learning PAM-L2 (Best and Tyler, 2007), propose that L2 learners’ ability to perceive L2 Frontiers in Communication | www.frontiersin.org 1 June 2021 | Volume 6 | Article 660917 Cebrian et al. When the Easy Becomes Difficult contrasts depends on the degree to which L2 phones are heard as, Flege (1992) also found that increased experience (i.e., longer or assimilated to, an exemplar of an L1 category. The model length of residence in the US) resulted in a more accurate describes several ways in which non-native or L2 sound categories production of the new vowel /æ/ by German speakers of L2 can be assimilated to L1 categories and predicts discrimination English, while the more similar English vowels /iː, ɪ, ε/ were ability of non-native (PAM) or L2 phones (PAM-L2), accordingly. equally accurately produced by all learners. Thus, for instance, if two non-native phones are perceptually The studies reviewed above generally support the SLM’s mapped onto two separate L1 categories (two category prediction that increased L2 experience, particularly when assimilation type), their discrimination will be easier than if the experience involves exposure to authentic L2 input and two phones are perceived as exemplars of the same L1 category predominant L2 use (Flege and Bohn, 2021, SLM-r), results in (Best, 1995). In turn, if the two phones perceived as belonging to more authentic categories for target sounds without a clear match the same L1 category differ in how closely they match the L1 in the L1. By contrast, L2 phones that are readily assimilated to L1 category (category goodness assimilation type), discrimination will phones may be less accurately categorized, and may not show be better than if the two phones are perceived as equal matches to improvement over time. However, there are two types of cases the L1 category (single category assimilation type). Other that do not follow these predictions. First, the case of target assimilation types involve combinations of categorized and phones that are similar to L1 phones and are nevertheless uncategorized sounds, whose discrimination accuracy may vary accurately produced or perceived. For instance, Fullana-Rivera depending on the degree of cross-language assimilation overlap and MacKay (2003) tested the effect of starting age of learning (Levy, 2009a), or perceived phonological overlap (Faris et al., and years of foreign language instruction on Catalan/Spanish 2018), that is, the overlap in the categorizations to native bilinguals’ production of English vowels. They observed that phones (see Tyler, 2021 for a description of all types and their English /iː/, with a clear counterpart in the L1, was always predictions). produced more accurately than the more dissimilar /ɪ/, The Speech Learning Model (SLM, Flege, 1995; Flege, 2003; although an improvement with /ɪ/ was observed with more see Flege and Bohn, 2021 for its recently revised version SLM-r) years of learning, consistent with the SLM’s expectations. proposes that the greater the perceived dissimilarity between L1 Further, in a study involving 240 Italian immigrants in and L2 phones, the greater the likelihood that learners establish Canada, Munro et al. (1996) found that degree of L2 English target-like categories. Like the PAM, the SLM proposes that the accentedness increased with age of arrival to the target language ability to discern between L1 and L2 categories is expected to country (AOA), which varied from 3.1 to 21.5 years. However, the improve with increased exposure to target-language input. English vowel /iː/, judged on the basis of an acoustic comparison to Therefore, phones that are perceived as different from any L1 be similar to Italian /i/, was more accurately produced by the phone may eventually be categorized more accurately than those Italian L2 speakers than more dissimilar vowels like /ʌ/ and /ɝ/. that are readily mapped onto L1 categories and are consequently The second type of outcome that does not follow from the perceived and produced in terms of that L1 category (Flege, 1995; predictions is the case of pairs of target phones involving a new Flege and Bohn, 2021). Still, some target language phones may in and a similar phone that obtain comparable results. For example, fact be sufficiently close to L1 counterparts that their perception Munro et al. (1996) also report that English /iː/ and /æ/, considered and production in terms of the L1 category may be undetected by similar and dissimilar to L1 categories, respectively, were native listeners (Flege, 1992). These can be referred to as identical comparably accurately produced. Flege et al. (1997) tested the or near identical L1-L2 pairs and would not require a separate L2 production and perception of L2 speakers from four L1 category (e.g., Spanish and English [f]). backgrounds and found that two groups of L1 Spanish L2 The prediction that target phones with no clear match in the English speakers differing in length of residence in the L1 will be learned more accurately than target phones with a United States (9 years vs. less than 6 months) produced the counterpart in the L1 is supported by some studies. Busà (1992) similar English vowel /ε/ more accurately than the new vowel examined the production of English vowels by Italians living in /æ/ (91–99% vs. 70–73% correct identification of the intended the United States and found that most of them produced English vowel by native English listeners). In addition, their scores for /uː/, classified as similar to Italian /u/, with lower F2 values than English /iː/ and /ɪ/, classified on the basis of acoustic measurements native English speakers did (that is, closer to Italian /u/). By as similar and new, respectively, were comparable (57–69% for /iː/ contrast, a larger number of Italians produced the new L2 vowel and 51–61% for /ɪ/). The question that remains is why similarly /ʊ/ more accurately than the similar vowel /u/. Similarly, research classified target phones (in terms of their similarity to L1 has found that Japanese learners of English show greater categories) are not always equally learned, while differently improvement in their perception and production of English /r/ classified target phones sometimes show comparable results. than of English /l/, where the former is more dissimilar from The inconsistencies found in some studies thus underscore the Japanese /r/ than the latter (Bradlow et al., 1997; Aoyama et al., need for a consistent and reliable method of measuring cross- 2004). Regarding the effect of L2 experience on the categorization linguistic similarity. One of the most common approaches of dissimilar phones, Jun and Cowie (1994) observed that involves the use of perceptual cross-language mapping tasks experienced Korean learners of English with a length of (Best, 1995; Flege et al., 1997; Strange et al., 2009; Tyler et al., residence of 16–31 years in the United States produced the 2014, among many others). In these tasks, commonly referred to new vowel /ɪ/ more accurately than inexperienced learners as categorization tasks or perceptual assimilation tasks (PAT), with a length of residence of 1.3–5.3 years. Further, Bohn and listeners are presented with non-native or L2 speech stimuli, and Frontiers in Communication | www.frontiersin.org 2 June 2021 | Volume 6 | Article 660917 Cebrian et al. When the Easy Becomes Difficult asked to 1) indicate to which L1 phonetic category each token is English /iː/-/ɪ/ contrast to a greater extent than native English most similar, and 2) rate its "goodness of fit" as an exemplar of speakers, who appear to rely mostly on spectral cues (Hillenbrand that category (see Materials and Methods section below). Tyler et al., 2000; Escudero and Boersma, 2004; Cebrian, 2006). This is (2021) argues that while there are some limitations with these found with learners whose L1 has temporal contrasts (e.g., /iː/-/i/ tasks, PATs are currently the most suitable method for measuring contrast), such as Japanese and Finnish (Ylinen et al., 2009; Grenon cross-linguistic similarity given that they evaluate different et al., 2019), and importantly also with speakers whose L1 has no possible sources of information that listeners attend to when vowel duration contrast, such as Mandarin Chinese, Korean, perceiving non-native sounds. This measure of perceived Russian, Catalan, Portuguese and Spanish (Bohn, 1995; Flege similarity has been found to be a good predictor of L2 et al., 1997; Wang and Munro, 1999; Escudero and Boersma, learning difficulty (e.g., Best and Strange, 1992; Guion et al., 2004; Cebrian, 2006; Mora and Fullana, 2007; Kondaurova and 2000; Strange et al., 2011; Tyler et al., 2014, among others). Francis, 2008; Morrison, 2008; Aliaga-García, 2011; Kivistö de The current paper aims to explore the link between similarity Souza et al., 2017). These results lend support to Bohn’s (1995) relations and L2 performance further by examining the perception desensitization hypothesis, which claims that L2 learners may not and production of the English /iː/-/ɪ/ contrast by a group of L1 be sensitive to L2 spectral distinctions that are not exploited in their Spanish/Catalan speakers. Previous studies have found that L1 L1, to which they have become desensitized. In these cases, Spanish/Catalan speakers have difficulty with the English /iː/-/ɪ/ duration may be used to differentiate L2 pairs with no parallel contrast but have not been able to relate it to the degree to which L1 spectral distinction. Further support comes from studies that each vowel is assimilated to a L1 category. Cebrian (2006) found show that learners exploit duration for specific contrasts only. Flege that two groups of L2 English speakers (20 Catalan speakers living and Bohn (1989) found that Spanish learners exploited temporal in Canada and 30 Catalan undergraduate students of English living cues to a greater extent than spectral cues in their identification of in Spain) perceptually assimilated Canadian English /iː, ε, eɪ/ to English /iː/ and /ɪ/, which do not have two clear L1 counterparts, Catalan /iː, ε, ei/, respectively, whereas English /ɪ/ obtained lower but overreliance on duration was not evident with /ε/ and /æ/, assimilation scores and goodness ratings as Catalan /e/. Perception which may be identified in terms of Spanish /e/ and /a/. Hence, of L2 sounds was examined with an identification test involving a duration is used to distinguish the pair of target vowels for which synthetic continuum from /iː/ to /ɪ/ to /ε/ varying in vowel quality the L1 does not have a comparable spectral contrast. Kondaurova and vowel duration. The two groups performed like native speakers and Francis (2008) conclude that several factors play a role in the in their perception of English /ε/, but differed from native speakers categorization of the /iː/-/ɪ/ contrast as a mostly temporal contrast, in their perception of English /iː/ and /ɪ/, which showed a including the role of duration as a phonetic cue in the learners’ L1 predominant reliance on temporal cues, unlike native Canadian (e.g., as a cue to stress differences), experience with different target- English speakers who relied mostly on spectral differences. These language varieties, psychoacoustic explanations based on the results do not follow from the results of the PAT since two English desensitization to spectral differences not exploited in the L1, vowels that were strongly assimilated to Catalan vowels, English /iː/ and a possible salience of duration and consequent ease of and /ε/, obtained different results, and the former patterned like the learning (Holt and Lotto, 2006; Hacquard et al., 2007; Kivistö- weakly assimilated vowel /ɪ/. The production of the target English de Souza et al., 2017). vowels by the 30 Catalan undergraduate students of English was This study examines the perception and production of the examined in a subsequent study (Cebrian, 2007). Data from eight English /iː/-/ɪ/ contrast by Spanish/Catalan bilinguals in relation native English speaker judges indicated that both English /eɪ/ and to their perceived similarity to L1 categories. The English high /ε/ were accurately produced (correctly identified 99 and 86% of front tense-lax pair /iː/-/ɪ/ has received a lot of attention in the the time, with goodness ratings reaching 5.5 and 5.2 out of 7, literature both because it poses a problem to learners from respectively). By contrast, English /iː/ and /ɪ/ obtained similarly different language backgrounds and also because it has a very lower rates of correct identification (73 and 71%, respectively), and high functional load in English (with more than 450 minimal /iː/ obtained higher goodness ratings than /ɪ/ (4.5 vs. 3.9) but still pairs according to Higgins, 2018). The fact that a large number of lower than /eɪ/ and /ε/. Acoustic measurements confirmed the these minimal pairs involve high-frequency words, and that many native speaker rating data and showed that the Catalan learners share the same grammatical category (e.g. feel-fill, leave-live), produced English /ε/ distinctly from English /iː/ and /ɪ/ and contribute to the high functional load of this contrast [Munro, produced the latter two with a large amount of spectral overlap 2021 (this volume)]. Thus, failure to distinguish these vowels can in terms of first and second formant. The learners, however, result in intelligibility problems. The perception and production produced a temporal difference between the tense and the lax of English /ε/ and non-rhoticized /ɜː/ (as in bed and bird in vowel (/iː/: 243 ms, /ɪ/: 153 ms), resulting in a ratio (1.62) that falls Southern British English, respectively) will also be analyzed for within the values reported for native English speakers (Hillenbrand comparison purposes as these vowels represent a clear case of a et al., 2000). similar and a dissimilar vowel. Table 1 shows the assimilation The inability to produce a spectral difference between English scores (percent identification in terms of Spanish vowel /iː/ and /ɪ/ thus seems to be linked at least in part to the fact that categories) reported for English /iː, ɪ, ε, ɜː/ in a few previous Catalan and Spanish learners of English tend to base the /iː/-/ɪ/ studies involving Spanish speakers. Generally, English /iː/ and /ε/ distinction on a temporal difference (e.g., Kondaurova and Francis, seem to be strongly assimilated to an L1 category (/i/ and /e/, 2008; Cerviño and Mora, 2009). In fact, L2 English learners have respectively), while /ɪ/ is more weakly assimilated to Spanish /e/. been found to exploit temporal cues in their categorization of the Cebrian (2019) also examined English /ɜː/, which patterned with Frontiers in Communication | www.frontiersin.org 3 June 2021 | Volume 6 | Article 660917 Cebrian et al. When the Easy Becomes Difficult TABLE 1 | Percent assimilation of English / iː, ɪ, ε, ɜː/ to Spanish vowels reported in recent studies (Goodness of fit ratings are given in parentheses when available). SSBE, GA (General American) and Can. Eng. (Canadian English) indicate the English variety of the stimuli. English stimuli L1 Spanish Cebrian (2019),a Baigorri et al. Morrison (2012),c Escudero and Flege (1991),e GA response SSBE (2019),b GA Can. Eng. Chládková (2010),d SSBE GA /iː/ /i/ 93 (5.9/7) 95–96 (8/9) 98–99 100 99 84–94 /ɪ/ /e/ 66 (5.3/7) 50 (3–6/9) 95–96 35 90 19–39 /i/ 33 (5/7) 49 3 21 10 36–68 /ε/ /e/ 97 (5.3/7) 87–94 (6/9) 95–96 77 96 44–81 /ɜː/ /e/ 64 (3/7) /a/ 14 (3.4/7) /o/ 13 (3.9/7) a Monolingual European Spanish speakers. b Late and early L2 English learners. c Monolingual European Spanish and Mexican Spanish speakers. d Monolingual Peruvian Spanish speakers. e L2 English speakers and monolingual Spanish speakers. /ɪ/ being weakly assimilated with Spanish /e/, with comparatively (Strange et al., 2009; Gilichinskaya and Strange, 2010; Tyler et al., low goodness of fit ratings.1 2014; Faris et al., 2018). For example, an individual analysis of the Another reason for the lack of a consistent link between cross- perceptual assimilation data collected by Cebrian (2019) showed linguistic perceived similarity and L2 performance may stem from that 15 out of the 29 Spanish speakers mapped English /ɪ/ to the amount of individual variation potentially found in L2 Spanish /e/ 75% of the time or more, eight participants chose performance. For instance, Munro et al. (1996) study involving Spanish /i/ as the closest L1 vowel at a similar rate, and six 240 Italian L2 English speakers reported a lack of consistent results participants divided their responses between Spanish /i/ and /e/ per vowel across L2 speakers as not all L2 speakers of similar (67% or less for each vowel). Thus, a fair amount of variability was experience produced the same vowels accurately. In a study on found among speakers of the same L1 background. Cantonese speakers’ production of English tense and lax high Therefore, another goal of the current study is to investigate the vowels, Munro (2021 (this volume)) also underscores the high relationship between perceived similarity and likelihood of degree of inter-speaker variability, which could not be related to accurate L2 category creation at the individual level. The cross- specific linguistic (word, phonetic context) or individual (language linguistic perceived similarity judgements of each individual are use, length of residence) factors. Still, the majority of studies make compared to their ability to produce and perceive the two vowels predictions about L2 performance for a specific group of learners distinctly. The study thus examines if previous findings that based on the average similarity judgements obtained for that English /iː/ and /ɪ/ are perceived and produced with similar group’s performance (e.g., Cebrian, 2006; Rallo Fabra and degrees of accuracy, despite differing in the extent to which Romero, 2012; Baigorri et al., 2019). This assumes that group they are mapped onto L1 vowels, are related to individual tendencies are representative of individuals’ perceptual differences in cross-linguistic perception, or to other factors judgements. Flege and Bohn’s (2021) SLM-r claims that (e.g., the relevance of different acoustic cues, the role of individual differences in L1 category precision may influence language instruction or the influence of orthography). The focus the discernment of cross-linguistic phonetic differences. In of this paper is the English high front tense-lax vowel contrast addition, Flege and Bohn caution that individuals may differ in (/iː/-/ɪ/), and two additional vowels are also examined, namely a the way they map L2 sounds onto L1 categories particularly in a vowel often reported to be strongly assimilated to an L1 vowel PAT. This is because listeners’ ratings may be affected by the (English /ε/) and one that lacks a consistent counterpart in the L1 interval between accessing long term memories of their native (English /ɜː/) in order to understand what is specific about the sounds in order to label the auditory stimulus and judging the /iː/-/ɪ/ contrast and what is determined by perceived similarity degree of similarity between that stimulus and the selected native relationships. category. In fact, studies have shown evidence of variability among individuals sharing the same L1 in perceived similarity judgments MATERIALS AND METHODS 1 Similar results have been obtained for similarity measurements involving Catalan A group of L2 English speakers performed a series of tests and English despite differences in vowel inventory between the two languages including a perceptual assimilation task, a vowel identification (Catalan has seven vowel phonemes (/i e ε a ɔ o u/, plus [ə] in unstressed position) task, a vowel discrimination task and a picture naming production while Spanish has five (/i e a o u/). For instance, Cebrian (2006) found that English /iː/ was assimilated to Catalan /i/ (99%, GR: 6.2/7), while /ɪ/ was split between task. The data here presented are part of a larger study evaluating Catalan /e/ (66%, GR: 3.5), /ε/ (20%, GR: 3/7) and /i/ (14%, GR: 2.5). See also the perception and production of a number of English vowels. This Cebrian (2021). paper focuses on the English /iː/-/ɪ/ and /ε/-/ɜː/ contrasts. Frontiers in Communication | www.frontiersin.org 4 June 2021 | Volume 6 | Article 660917 Cebrian et al. When the Easy Becomes Difficult Participants carried out in a soundproof booth at the speech laboratory at The participants were 43 Catalan/Spanish bilinguals (37 females, University College London, in London, United Kingdom, and mean age 19.2, standard deviation 1 year) who were first-year were digitized at a 44.1 kHz-sampling rate and normalized for English Studies undergraduate students at Universitat Autònoma intensity [70 dB in sound pressure level (SPL)]. The best tokens de Barcelona (UAB), Barcelona, Spain. They reported having per talker were selected, based on auditory judgements and started learning English in school before the age of 12 (the spectrographic analysis. Neighboring sounds have been found to average starting age of learning was six; average length of affect perceptual assimilation judgements (e.g., Bohn and Steinlen, learning 13 years), and were at the time attending classes in 2003; Levy, 2009b). Hence stimuli were edited to include from the English at university between 6 and 15 h per week. Most release of the /b/ until the closure of /t/, thus eliminating potential participants reported speaking English at university (with cross-linguistic and individual differences in stop production classmates, foreign students, and teachers) at least half the time, (prevoicing of /b/, /t/ release), while maintaining intact the cues although they used English less often with family, friends and at to the vowel. work. The majority had not lived in an English-speaking country and those who had, had spent less than 6 months (an average of Perceptual Assimilation Task 2 weeks), generally as part of a summer stay. Participants reported In the perceptual assimilation tasks, listeners were required to having no hearing impairments. All 43 participants performed the identify each English vowel stimulus in terms of one of several three perception tasks, but five participants did not complete the possible L1 categories by clicking on one of the response options production task and, thus, only the production of the remaining 38 presented on a computer screen. Upon selecting an L1 category, participants was analyzed (33 females, mean age 19.4, standard listeners provided a goodness of fit rating on a 7-point scale, deviation 0.9 years). The participants were all native speakers of where 1 meant a poor example of the selected vowel and 7 meant Spanish although some spoke Catalan as well. Participants a good, native-like example. The response options consisted of completed an online personal and language background the most common and unequivocal spelling for each Spanish questionnaire based on the Bilingual Language Profile (BLP; vowel and diphthong () together with a Birdsong et al., 2012). Among other information, the BLP monosyllabic word illustrating that vowel, namely di (for /i/), se renders a score that ranges from 218 to −218 representing the (/e/), da (/a/), do (/o/), tu (/u/), hay (/ai ̯/), rey (/ei ̯/), hoy (/oi ̯/), two monolingual endpoints, with values close to zero indicating meaning say, self (reflexive pronoun), give, do (musical note), balanced bilingualism. According to the questionnaire, 22 you, there is, and today, respectively. Every vowel appeared in 12 participants were Spanish-dominant, 11 appeared to be balanced trials (3 talkers × 2 tokens × 2 repetitions). This paper reports the in Spanish and Catalan, and 10 were more dominant in Catalan. results for the vowels under study, namely /iː, ɪ, ε, ɜː/. See Cebrian This difference among participants need not be problematic as one (2019) for a study covering a near complete set of SSBE vowels. of the goals of this paper is precisely to examine how individual variation in cross-linguistic perception may affect L2 performance. Nevertheless, an analysis of the Catalan-dominant bilinguals’ Vowel Identification and Vowel results in the PAT revealed that their responses for the vowels Discrimination Tasks under study did not differ significantly from the remaining Spanish Stimuli speakers’ performance (see Perceptual Assimilation Task). All tests The stimuli used in the identification and discrimination tasks took place at the Speech Laboratory at UAB on two different days. consisted of high-frequency monosyllabic English words. Each Participants performed the production task and the perceptual word shared the initial and final consonants with at least one assimilation task in one session and the identification and other word, thus resulting in a series of minimal pairs (e.g., bead, discrimination tasks in a second session. bid, bed, bird). Half the words ended in a voiced stop and the other half ended in a voiceless stop. The words were produced by two Cross-Language Perception native speakers of SSBE, namely a 23 year-old female and a 33 year- Stimuli old male, who were different from the speakers who produced the The stimuli used in the perceptual assimilation task (PAT) were a stimuli for the PAT. The speakers were recorded in a soundproof subset of the stimuli used in a previous study (Cebrian, 2019) booth at the speech laboratory at University College London. The and consisted of the Standard Southern British English (SSBE) recordings were digitized at 44.1 kHz sampling rate and normalized vowels /iː, ɪ, ε, eɪ, aɪ, æ, ɑː, ɜː, ʌ/ produced in /bVt/ words.2 for intensity (70 dB in SPL). The same stimuli were used in the These words were elicited from three monolingual male speakers identification task, presented individually, and the discrimination of SSBE (mean age 35), who had spent most of their lives in the task, arranged in pairs (e.g., bead-bid, bead-bead), as explained South of England and were living in London. The recordings were below. These tasks were part of a larger study examining the perception of seven English vowels (/iː, ɪ, ε, æ, ɑː, ɜː, ʌ/). 2 Identification Task Participants were exposed to different English varieties through music, TV and cinema, and the internet (according to the language background questionnaire, their Participants were asked to identify the vowel in the stimulus using average exposure to British and American English was 54 and 46%, respectively). Still, one of seven response options appearing on a computer screen. British English is the main variety taught in schools and language centers in Spain The options consisted of a phonetic symbol together with two and it is the variety described in most textbooks used at UAB. common words representing each sound, namely /æ/ ash, mass; /ʌ/ Frontiers in Communication | www.frontiersin.org 5 June 2021 | Volume 6 | Article 660917 Cebrian et al. When the Easy Becomes Difficult sun, thus; /ɪ/ fish, his; /iː/ cheese, leaf; /ɜː/ earth, first; /ε/ less, west; ends with a correlational analysis contrasting the results of the different /ɑː/ arm, palm. There were eight stimuli per vowel (four words tasks. Each set of results is followed by a brief interim discussion. produced by two different speakers) and each stimulus appeared twice throughout the task, resulting in a total of 16 trials per vowel. Perceptual Assimilation Task Perceptual Assimilation Tasks Results Discrimination Task The percentage of times participants identified each target vowel The discrimination task was a categorical AX same/different as one of the Spanish responses (i.e., assimilation scores) and the discrimination task, including several pairs of English vowels. corresponding average goodness of fit ratings (GR) were The data presented in this paper focuses on the /iː/-/ɪ/ and /ε/-/ɜː/ calculated. In addition, following Guion et al. (2000), a fit pairs. For each pair, several stimuli were created by including the index score (FI) was calculated by multiplying the two possible speaker orders (speaker 1-speaker 2, speaker 2- identification proportion by the goodness of fit rating. This speaker 1) and vowel orders (e.g., /iː/-/ɪ/ and /ɪ/-/iː/). Half the composite score is meant to distinguish between cases of stimuli included the same vowel category (same-category trials, comparable assimilation percentages that differ in GR, and /iː/-/iː/), and half were different-category trials (/iː/-/ɪ/). This was used by Guion et al. (2000) to determine the likelihood resulted in a total of 16 different-category trials per vowel pair, of accurate discrimination for pairs of L2 vowels. Further, given and 8 same-category trials per vowel. The interstimulus interval that individual differences in perceived similarity are often was 1.15 s so that participants used phonetic information stored observed, the number of participants who selected a given in long term memory instead of relying on sensory memory (e.g. response at a given level of consistency was tallied in order Højen and Flege, 2006). Participants responded by clicking on the to assess the degree of agreement among listeners (Strange et al., words “same” or “different” displayed on a computer screen. 2009; Gilichinskaya and Strange, 2010; Cebrian, 2021). To that effect, the categorization thresholds used in some previous Production Task studies were used, i.e., 70% (Tyler et al., 2014) and 50% Thirty-eight of the 43 participants who performed the PAT and the (Bundgaard-Nielsen et al., 2011; Faris et al., 2018). The identification and discrimination tasks also completed a picture results are presented in Table 2. naming task in which they were asked to name 44 different pictures Recall from Section Participants that many of the twice. The list included the four words containing each target vowel. participants spoke Catalan in addition to Spanish and Specifically, the words were bed, bet, head, pet for /ε/, bird, heard, participants differed in how strongly dominant in Spanish hurt, dirt for /ɜː/, bid, bit, dip, lid for /ɪ/, feed, feet, league, leak, for /iː/. they were. An analysis was conducted to see if responses in The words were presented in a random order, which was the same the PAT were influenced by the bilinguals’ language dominance. for all participants. The first four words were fillers used to The results for the subgroup with a clear dominance in Spanish familiarize participants with the task. For each word, a picture (n  22) were compared to those who showed a balanced was displayed on a computer screen (MS PowerPoint was used). In dominance (n  11) or a more Catalan dominance (n  10). order to ensure that the right word was produced, the target word The results for English /ε, iː, ɪ/ were practically identical across was shown at the bottom of the screen in written form. The written groups, while in the case of English /ɜː/, a greater number of word disappeared after two seconds, then participants were Spanish /a/ responses were obtained as dominance in Catalan instructed to name the picture twice, prompted by the increased, reaching 20%. A series of Mann-Whitney U-tests appearance of a dialogue balloon next to the picture. Participants yielded no significant effect of language dominance on were recorded in a soundproof chamber at the speech laboratory at assimilation scores, confirming that the general pattern of UAB, Spain. In addition, the production of 13 native English results was very comparable across dominance groups. speakers (seven female) was also collected using the same task to In terms of individual variation, out of the 43 participants provide native English values to compare the L2 production to. These who performed the PAT, 23 assimilated English /ɪ/ to Spanish were native speakers of Southern British English in their twenties and /e/ at least 70% of the time, six learners chose Spanish /i/ as the early thirties. The group consisted of a mixture of international best L1 match 70% of the time or more, and the remaining 14 students and English teachers residing in Barcelona (for 2–10 years; learners chose either option between 25 and 67% of the time. reported weekly use of English: 75% of the time), who were recorded This shows that, though the predominant pattern was to at the speech laboratory at UAB, and five undergraduate students at perceive English /ɪ/ as most similar to Spanish /e/, there was Queen Mary University, London (monolingual English speakers, a certain degree of variability among participants. Responses three of whom had taken Spanish lessons but barely used Spanish for English /ε/ and /iː/ were very consistent, with all outside the classroom), recorded in a sound attenuated room at that participants identifying English /ε/ with Spanish /e/ 70% of institution. Recordings were digitized at 44.1 kHz sampling rate and the time or more, and almost all participants (39) assimilating normalized for intensity (70 dB in SPL). English /iː/ consistently to Spanish /i/, while four yielded responses split between Spanish /i/ and /ei/. Regarding /ɜː/, 28 participants chose Spanish /e/ as the closest vowel at least RESULTS AND INTERIM DISCUSSIONS 70% of the time, three participants selected Spanish /a/, two other participants selected Spanish /o/ and the remaining 10 This section presents the results of the perceptual similarity task, participants did not choose a specific L1 vowel more than 67% followed by the L2 perception and L2 production results. The section of the time. Frontiers in Communication | www.frontiersin.org 6 June 2021 | Volume 6 | Article 660917 Cebrian et al. When the Easy Becomes Difficult TABLE 2 | Perceived similarity between nonnative (English) vowels and L1 (Spanish) vowels. (Num. of Ss assim. ≥70%/50%  number of subjects who selected a given response 70%/50% of the time or more). The total number of participants was 43. English stimuli L1 Spanish % Assimilation Goodness Rating Fit Index Num. of Num. of response Ss assimn. Ss assimn. ≥70% ≥50% /iː/ /i/ 94 6.5 6.1 39 43 /ei ̯/ 5.8 4.4 0.3 1 /ɪ/ /e/ 67 5.9 4.0 23 32 /i/ 32 5.4 1.6 6 12 /ε/ /e/ 99 5.3 5.2 43 43 /ɜː/ /e/ 72 4.3 3.1 28 34 /a/ 15 2.8 0.4 3 4 /o/ 11 4.3 0.5 2 5 Discussion of Perceptual Assimilation Tasks Results 5.3 vs. 3.1). PAM-L2 predicts that discrimination of these two types Assuming a categorization threshold of 70% (Tyler et al., 2014), of contrasts by L2 speakers will be more difficult than between English /iː/, /ε/ and /ɜː/ were categorized as Spanish /i/, /e/ and /e/, pairs that assimilate to two different non-native vowels (e.g., respectively. English /ɪ/ did not reach the categorization threshold English /iː/ and /ε/) and their discrimination accuracy may of 70%, but was assimilated above chance to more than one L1 depend on the degree of perceived phonological overlap. Thus, phone (Spanish /e/ and /i/) and thus illustrates an uncategorized- we may expect that English /iː/-/ɪ/ will be more accurately clustered type of assimilation (Faris et al., 2018). The current discriminated than English /ε/-/ɜː/. results are on the whole in agreement with the previous studies In terms of the SLM (Flege, 1995; and its recently revised showing that Spanish speakers consistently assimilate English /iː/ version, the SLM-r; Flege and Bohn, 2021), all four English vowels and /ε/ to Spanish /i/ and /e/, and that English /ɪ/ displays a less will be initially categorized in terms of the closest L1 phones. consistent pattern (Escudero and Chládková, 2010; Morrison, Exposure to authentic target language input may enhance the 2012; Baigorri et al., 2019; Cebrian, 2019, see Table 1 above). ability to differentiate the L2 from L1 sounds and thus establish In fact, the results closely replicate those obtained by Cebrian separate categories for the L2 sounds. This is more likely for (2019), which involved a near-complete set of SSBE and Spanish English /ɪ/ and /ɜː/, judged to be more different from L1 vowels and diphthongs and tested a group of Spanish speakers with categories, than for English /ε/ and /iː/, which have a clear little or no knowledge of English. The only difference is that English match in the L1. In fact, the latter two may pattern as near- /ɜː/ fell short of a 70% categorization threshold in the previous study identical to L1 categories. Flege (1992) suggests that in order to (64% as Spanish /e/ vs. 72% in the current study) and was thus know if a non-native sound is perceived to be indistinguishable classified as uncategorized-clustered. Setting the categorization from a native category, this non-native phone should be threshold at 70% (or 50%) is an arbitrary decision, as there are undetectable when produced in a native context. In brief, no clear criteria supporting one or another cut-off (see Tyler, 2021, English /iː/ and /ε/ pattern as highly assimilated to Spanish /i/ for discussion). In fact, in both studies English /ɪ/ obtained a higher and /e/, while English /ɪ/ and /ɜː/ are weakly assimilated. The GR as Spanish /e/ than English /ɜː/ did, indicating that the latter was prediction is that the level of accuracy in the perception and perceived as a more dissimilar vowel. The lower GRs for /ɜː/ may production of English /iː/ and /ε/ by Spanish L2 speakers will be also be related to the greater differences across English varieties similarly high, while for /ɪ/ and /ɜː/ performance may be regarding this vowel, which may have affected listeners’ familiarity originally worse and accuracy will depend on the extent to with SSBE /ɜː/. which learners detect differences between L1 and L2 sounds Considering both the degree of categorization and the FI, what thanks to continuous authentic input. the data show is that English /ε/ and /iː/ are strongly assimilated to Spanish /e/ and /i/, while English /ɪ/ and /ɜː/ show weaker degrees L2 Perception of assimilation to Spanish /e/. The main focus of the current study Identification Results is the English /iː/-/ɪ/ contrast. The PAT results show that the two The results of the identification task are presented in Figure 1 and English vowels differ notably in degree of assimilation to an L1 Table 3. Figure 1 shows the median and distribution of the category (FIs of 6.1 vs. 4, respectively). The pair patterns as an percent correct identification for each of the four target vowels. uncategorized-categorized type of assimilation in PAM-L2’s terms Table 3 shows the average correct identification and (as Spanish /i/ and /e/, respectively), with a cross-language misidentification responses (scores below 3% are omitted). As assimilation or perceived phonological overlap (Levy, 2009a; can be observed, English /ε/ was the most accurately identified, Faris et al., 2018) of 32% as Spanish /i/. The English /ε/-/ɜː/ followed by /ɪ/, while identification was worse with /iː/ and /ɜː/. contrast patterns as a category-goodness assimilation type, with a The identification results were submitted to a series of generalized considerable perceived phonological overlap (72%) and a notable linear mixed models (GLMM) with score (correctly or incorrectly difference in degree of assimilation to Spanish /e/, as English /ε/ is identified) as the dependent variable. The best fitting model was a perceived as a better match for Spanish /e/ than English /ɜː/ (FIs: GLMM with Target vowel as fixed factor, and participant as Frontiers in Communication | www.frontiersin.org 7 June 2021 | Volume 6 | Article 660917 Cebrian et al. When the Easy Becomes Difficult FIGURE 2 | Boxplots of discrimination accuracy for the two English FIGURE 1 | Boxplots of identification accuracy for each of the four target vowel contrasts examined. Boxplots represent the quartile ranges of the English vowels. Boxplots represent the quartile ranges of the scores (top of scores (top of box: 75th percentile, bottom of box; 25th percentile; line in the box: 75th percentile, bottom of box; 25th percentile; line in the middle: middle: median). The whiskers represent the highest and lowest values median). The whiskers represent the highest and lowest values that are that are not outliers or extreme values, which are indicated by circles outside not outliers or extreme values, which are indicated by circles outside the the whiskers. whiskers. TABLE 3 | Identification of the four English target vowels by 43 Spanish/Catalan Discrimination Results learners of English. Mean correct identification percentages are highlighted in The results of the discrimination task are presented in Figure 2. bold, misidentifications equal to or lower than 3% have been omitted. The /ε/-/ɜː/ pair was discriminated more successfully than the English target vowels /iː/-/ɪ/ pair, with mean percentages of correct discrimination reaching 87% for /ε/-/ɜː/ and 66% for /iː/-/ɪ/. The results of a Responses /iː/ /ɪ/ /ε/ /ɜː/ GLMM with Vowel pair as fixed effect and a random slope for /iː/ 58 21 participant3 confirmed that the difference between the two vowel /ɪ/ 39 72 pairs was significant (F (1, 1,030)  38.15, p < 0.001). /ε/ 4 90 4 /ɜː/ 7 47 Discussion of L2 Perception Results /ʌ/ 23 The finding that /ε/ was very successfully identified and /ɜː/ was /ɑː/ 19 /æ/ 4 poorly identified can be explained in terms of the PAT results, as the former was strongly assimilated to an L1 category, while the latter barely reached a categorization threshold of 70% (99 and 72%, respectively, as Spanish /e/). Two studies testing the effect of random intercept.3 The results showed a significant effect of high variability perceptual training on SSBE vowel identification Target vowel (F (3, 2,444)  65.3; p < 0.001). Bonferroni-adjusted by Catalan/Spanish learners of English (Carlet and Cebrian, pairwise comparisons showed that all between-vowel differences 2019) and Spanish monolingual learners of English (Fouz- were significant (p < 0.001). Regarding the pattern of González and Mompean, 2020) also found that /ɜː/ was the misidentifications, English /iː/ was most frequently misheard most poorly identified vowel, but its identification tended to as /ɪ/ and vice versa, while /ɜː/ was most often misidentified as improve the most as a result of increased L2 experience (namely /ɑː/ or /ʌ/. perceptual training), as expected for a dissimilar vowel (e.g., Flege, 1995). On the other hand, the fact that /ɪ/ was better 3 For the statistical analyses on the L2 perception and production data, several identified than /iː/ (72 vs. 58%, respectively) cannot be equally generalized linear mixed models (GLMM) were considered, which included the explained by the PAT results. The expectation would be for /iː/ to independent variables in each case as the fixed effects, and different combinations pattern with /ε/ and be successfully identified given that both of random intercepts and slopes for participant and stimulus. The selection of the English vowels were strongly assimilated to L1 vowels. Still, better best model was based on the lowest Akaike Information Criterion for a model that identification of /ɪ/ than of /iː/ has been reported for similar included participant as random intercept or random slope, given the theoretical populations (Cebrian and Carlet, 2014; Carlet and Cebrian, 2019; relevance of participant variability in the current study. In fact, in every case, the results for the different models were very consistent and the levels of significance Fouz-González and Mompean, 2020). for each factor and pairwise comparison were very similar across models. IBM When misidentified, SSBE /iː/ was usually perceived as /ɪ/. The Corp (2017) software was used. results of the PAT showed that /ɪ/ was assimilated to Spanish /i/ Frontiers in Communication | www.frontiersin.org 8 June 2021 | Volume 6 | Article 660917 Cebrian et al. When the Easy Becomes Difficult about a third of the time (32%), that is, there was a 32% overlap in perceptual assimilation patterns is analyzed in Cross-Task the assimilation of English /iː/ and /ɪ/ to Spanish /i/. This overlap, comparison. Correlational analysis. and the relatively high goodness of fit ratings (GR) obtained for In order to compare the L2 data to native English data, and English /ɪ/ as Spanish /i/ (5.4/7), indicate a degree of perceptual to group males and female speakers together, the formant confusion that may explain the misidentifications of English /iː/ values were normalized following the Lobanov method as /ɪ/. On the other hand, the pattern of misidentifications of /ɜː/, (Lobanov, 1971). A Lobanov transformation is a vowel mostly as /ɑː/ or /ʌ/ but hardly as /ε/, may be linked to the extrinsic normalization method that has been found to comparatively low GRs for /ɜː/ as Spanish /e/ (4.3). Thus, learners render superior normalization outcomes to other methods, tended to identify as English /ε/ only stimuli that were clear particularly if the data set examined includes a large set of exemplars of this vowel. In brief, the identification results are not vowels (Adank et al., 2004; Recasens and Espinosa, 2009). expected from the PAT results, as the two strongly assimilated Recall that, although this paper focuses on four vowels, the vowels (/iː/ and /ε/) obtained very different results, and the two current data is a subset of a larger set including a larger number weakly assimilated vowels (/ɪ/ and /ɜː/) also patterned very of vowels, which were taken into account in the normalization differently. procedure. Normalization was carried out using NORM Regarding the discrimination task, the results seem to show (Thomas and Kendall, 2007). Figure 4 displays the spectral more difficulty with the /iː/-/ɪ/ contrast than some previous characteristics of the native and L2 productions, averaged studies, which reported discrimination accuracy rates close to across the 38 L2 speakers (L2S; 33 females) and the 13 or over 75% or A’ scores denoting sensitivity to the contrast, native English speakers (NES; seven females). As can be i.e., over 0.6 (Mora and Fullana, 2007; Rallo Fabra and Romero, observed, as a group the L2 speakers seemed to produce 2012; Baigorri et al., 2019). Comparisons across studies are English /ɜː/ and /ε/ as two separate vowels and with values complicated, however, by methodological differences such as that were close to those of NES’s. By contrast, the learners’ the selection of English contrasts examined and the variety of production of /iː/ and /ɪ/ were much closer than the respective English tested (SSBE vs. American English). In terms of the PAM, productions by native speakers. /iː/-/ɪ/ was classified as a categorized-uncategorized clustered type A series of GLMMs were conducted3 with F1 and F2 of assimilation, with partial perceptual overlap. The English pair (Lobanov-normalized values) as the dependent variables to /ε/-/ɜː/ was classified as a category-goodness difference type of assess if the L2 speakers produced the English vowels assimilation (and as a categorized-uncategorized clustered type in differently and if the L2 production differed from NES a previous study, Cebrian, 2019). According to Faris et al. (2018), production. In both cases the best fitting model was a the two types of assimilation pose a similar degree of GLMM with Group (learners vs. NES) and Vowel (the four discrimination difficulty, and thus discrimination accuracy target vowels) as fixed effects, and random intercepts for depends on the amount of phonological overlap. The /ε/-/ɜː/ participant and for word. Regarding F1 values, the model pair, with a greater phonological overlap (72% as Spanish /e/), revealed a significant effect of Vowel (F (3, 1,801)  238.02, was expected to pose a greater difficulty than the /iː/-/ɪ/ contrast p < 0.001) and a significant Group by Vowel interaction (F (3, (32% overlap as /i/), but the results showed the opposite pattern. 1,801)  82.4, p < 0.001), but no significant effect of Group (F (1, 1,801)  2.488, p  0.115). Bonferroni-adjusted pairwise L2 Production comparisons indicated that the L2 speakers differed from the Production Results native speakers with respect to /ε/, /iː/ and /ɪ/ (p < 0.001), but not The vowel production of the 38 L2 speakers (recall that five to /ɜː/ (p  0.15), which explains the significant interaction. NES participants only completed the perception tasks) and 13 native produced significantly different F1 values for all four vowels (p < English speakers was analyzed acoustically in terms of first and 0.001), while L2S produced significant differences between all second formant (F1 and F2) and vowel duration. Vowel formants vowels (p < 0.001) except for the /iː/-/ɪ/ contrast (p  0.335). were measured from a 25 ms window located manually at a With respect to F2, all main effects and the interaction reached steady-state portion between one third and two fourths into significance (Vowel: F (3, 1,801)  235.92, p < 0.001; Group: F (1, the vowel, using Praat (Boersma and Weenink, 2018). 1801)  4.6, p  0.031; Group x Vowel: F (3, 1,801)  105.87, p < Figure 3 displays the mean F1 and F2 values in Hertz per 0.001). Pairwise comparisons in this case indicated that the L2 vowel for each of the L2 speakers, showing the data for female speakers differed from NES with respect to /ɜː/, /iː/ and /ɪ/ (p < and male L2 speakers separately. Each individual is represented 0.001), and /ε/ (p < 0.01). The smaller level of significance for the by a number. As can be observed from the distribution of the group comparison involving /ε/ may explain the significant individuals’ production, there is considerable spectral overlap interaction. NES produced significantly different F2 values between /iː/ and /ɪ/, while /ε/ and /ɜː/ appear to be distinguished for all four vowels (p < 0.001). As was found for F1, L2S more clearly. As expected, a fair amount of individual variation produced a significant F2 difference between all vowels (p < can be observed. For example, male speaker 15 produced 0.001) except for the /iː/-/ɪ/ contrast (p  0.08). In brief, although practically indistinguishable /iː/ and /ɪ/ (see Figure 3). By L2 productions generally differed from NES’s, the production of contrast, male speaker 47 produced these two vowels with very /ε/ and /ɜː/ by L2 speakers deviated to a lesser extent from NES’s different spectral values. The issue of whether this difference in production. In addition, learners produced significant spectral performance can be predicted from individual differences in differences between all vowels except for the /iː/ and /ɪ/ pair. Frontiers in Communication | www.frontiersin.org 9 June 2021 | Volume 6 | Article 660917 Cebrian et al. When the Easy Becomes Difficult FIGURE 3 | F1 x F2 values (in Hertz) of the four English vowels produced by female (left, N  33) and male (right; N  5) L2 English speakers. Numbers represent individual participants. FIGURE 4 | Normalized F1 x F2 values of each of the four vowels averaged across the native English speakers (NE) and the L2 English speakers (L2). The Lobanov transformation has been used (see text). Given previous findings about Spanish/Catalan speakers’ Bonferroni-adjusted pairwise comparisons indicated that the L2 reliance on temporal cues for the English /iː/-/ɪ/ contrast speakers differed from the native speakers in the duration of the (Cebrian, 2006; Rallo Fabra and Romero, 2012), vowel duration tense vowels /ɜː/ and /iː/ (p < 0.001), but not regarding the lax was also examined. Table 4 provides the means and standard vowels /ɪ/ (p  0.304) and /ε/ (p  0.249), which explains the deviations for each group and vowel, as well as the duration ratio significant interaction. Both groups produced highly significant for /iː/-/ɪ/ and /ɜː/-/ε/. As can be observed, both groups produced duration differences between all vowels (p < 0.001), except for the the tense vowels with greater duration than lax vowels, but the difference between /ε/-/ɪ/ for NES, which was significant at the p < duration difference was greater for NES. The results of a GLMM 0.05, and the difference between /ε/-/iː/, which was not significant exploring vowel duration with Group and Vowel as fixed effects for the L2 speakers (p  0.747). and a random slope for participants revealed a significant effect of It has been argued that perceptual measures such as native Vowel (F (3, 1,802)  469.25, p < 0.001) and a significant Group by speaker judgements are a more appropriate method for assessing Vowel interaction (F (3, 1,802)  58.41, p < 0.001), but no L2 production and accentedness than acoustic analyses (e.g., significant effect of Group (F (1, 1,802)  3.176, p  0.075). Munro, 2008; Derwing and Munro, 2015). A preliminary Frontiers in Communication | www.frontiersin.org 10 June 2021 | Volume 6 | Article 660917 Cebrian et al. When the Easy Becomes Difficult TABLE 4 | Mean vowel duration (in ms) and duration ratio for L2 speakers (L2S) Spanish/Catalan learners of English, whose /ɜː/ was judged to and English native speakers (NES). be comparatively accurate by native English judges. Raters in that L2S (n = 38) NES (n = 13) study also judged /iː/ to be better produced than /ɪ/, in line with our preliminary native speaker judgements. Similarly, Carlet and /iː/ 154 (38) 190 (29) Kivistö-de Souza (2018) found that the productions of /iː/ by /ɪ/ 140 (29) 129 (24) /ɜː/ 209 (38) 257 (35) Spanish/Catalan learners of English were judged to be more /ε/ 153 (28) 141 (24) accurate (5.1 on 9-point Likert scale) than those of /ɪ/ (4.8). /iː/-/ɪ/ ratio 1.11 (0.22) 1.50 (0.24) Overall, then, previous studies tend to find an accurate /ɜː/-/ε/ ratio 1.38 (0.18) 1.88 (0.33) production of English /ε/, and a spectral overlap in the production of /iː/ and /ɪ/, which are found to be less accurately produced than /ε/. perceptual analysis of the production data was carried out. Four Regarding vowel duration, the results of the current study judges (native speakers of SSBE aged between 25 and 36) rated a show a smaller reliance on temporal cues for the /iː/-/ɪ/ contrast selection of all the words produced by the L2 speakers and the than previous studies involving Catalan/Spanish L2 English native English controls (recall that the current data form part of a speakers (Cebrian, 2007; Aliaga-Garcia and Mora, 2009; Rallo larger study involving a greater number of vowels). On a given Fabra and Romero, 2012), as the tense vowel was only 11% longer trial raters first identified the vowel in the stimulus in terms of one than the lax vowel (compared to 50% for NES). Still, there was of ten English words presented on a computer screen (including variability among the participants. One of the SSBE native meet, sit, set and shirt) and then provided a goodness rating on a speakers did not produce a temporal difference between /iː/-/ɪ/. 7-point scale where 7 meant native-like. The results show that the Regarding the L2 speakers, eight had duration ratio values for L2 speakers’ production of English /ε/ was very accurate, with /iː/-/ɪ/ within the range of the SSBE speakers’ values, while 15 100% identification and an average goodness rating (GR) of 5.8 had ratio of 1 or lower. In order to see if there was an inverse out of 7. Vowel /ɜː/ received very high identification scores too, relationship between producing a temporal difference and a 89%, and a relatively high GR of 5. Vowel /iː/ was correctly spectral difference, the relationship between the /iː/-/ɪ/ duration identified 84% of the time with a GR of 4.9, and finally vowel /ɪ/ ratio and the spectral Euclidean distance between their was the least accurately produced, reaching only 50% correct production of /iː/ and /ɪ/ was examined. No significant identification, with a GR of 4.3, and being heard as /iː/ 43% of the correlation was found, indicating that those participants who time with a GR of 4.3. Although these results are preliminary as so relied more on spectral differences were not necessarily those far only two judgements per stimulus have been obtained, they who relied less on temporal differences. appear to confirm the acoustic analysis and point to the strongly In summary, the production of /ε/ follows predictions for a assimilated vowel /ε/ as the most accurately produced vowel, highly assimilated vowel that may pattern as a near identical followed by /ɜː/ and /iː/, while /ɪ/ was the least accurately vowel. Results for /ɜː/ are unexpected for a comparatively new produced. vowel that was poorly identified. Results for /iː/ and /ɪ/ show that, while /iː/ tends to be better produced, particularly as judged from Discussion of L2 Production Results a native perceptual perspective, it is not as accurately produced as The production results show better production of /ε/ and /ɜː/ /ε/ despite comparable assimilation scores for English /ε/ and /iː/ than of /iː/ and /ɪ/. These results are consistent with previous as Spanish /e/ and /i/, respectively. In fact, /iː/ patterns more with studies involving Spanish learners of English. Flege et al. (1997) /ɪ/ than with /ε/ in production accuracy. The analysis also reported that two groups of Spanish speakers differing in the revealed that learners made use of temporal cues, but to a amount of time spent in the United States (0.5 vs. 9 years) lesser extent than what has been reported in previous studies. produced English /ε/ very accurately (correctly identified between 91 and 99% of the time by native English judges), while /iː/ and /ɪ/ obtained similarly lower identification scores Cross-Task Comparison. Correlational (51–61 and 57–69%, respectively). Cebrian (2007) also found that Analysis English /ε/ was the most accurately produced vowel by a group of One of the goals of this paper is to examine if individual variation 30 Spanish/Catalan speakers of English (86% correct in the perceived similarity between L1 and L2 sounds is related to identification and a 5.3/7 goodness rating by native English perception and production accuracy, as would be predicted by the speakers), while /iː/ and /ɪ/ were less accurately produced (/iː/ SLM-r (Flege and Bohn, 2021). In order to answer this question, a 73% and GR 4.5/7; /ɪ/: 71%, GR: 3.9). The native speaker number of Spearman correlations were conducted relating cross- judgements were consistent with the acoustic analysis of L2 linguistic perceptual similarity measures and L2 perception and data, which revealed a large overlap in F1 and F2 space production measures. Specifically, the perceptual assimilation between /iː/ and /ɪ/, while /ε/ was more distinctly produced. measures included in the analysis were percent assimilation, Further, in a study involving Catalan learners of English, Rallo goodness ratings, and the composite fit index score for each Fabra and Romero (2012) reported that the learners produced /ɪ/ vowel. Identification accuracy for each L2 vowel and more accurately than /iː/ (71% vs. 49% correct identification by discrimination accuracy for each vowel contrast were the L2 native speakers). Regarding /ɜː/, the current results are in perception measures. Finally, two types of production measures agreement with Carlet and Cebrian’s (2019) study involving were explored. First, the extent to which speakers distinguish the Frontiers in Communication | www.frontiersin.org 11 June 2021 | Volume 6 | Article 660917 Cebrian et al. When the Easy Becomes Difficult two vowels in each contrast, that is, the Euclidean distances 2007; Kondaurova and Francis, 2008; Morrison, 2008; between /iː/ and /ɪ/, and between /ɜː/ and /ε/. In addition, to Morrison, 2009). L2 speech theories relate accurate explore the degree to which L2 speakers approximated native perception and production of target language phones to the production, the Euclidean distance between each L2 individual’s ability to discern differences between L1 and L2 sounds (Best, production of each vowel and the native English speakers’ average 1995; Flege, 1995; Escudero, 2005; Best and Tyler, 2007; Flege production was also examined. and Bohn, 2021). This study set out to examine if this difficulty Only a subset of the possible correlations reached significance is related to the degree of perceived similarity between L1 and (see Supplementary Material in the Appendix for plots of the L2 vowels, both in terms of group results and individual significant correlations). Accuracy in the identification of vowel judgements of cross-linguistic similarity. To that effect, a /iː/ was moderately correlated with the percent assimilation of group of 43 Catalan/Spanish bilinguals, who were undergraduate English /ɪ/ to Spanish /e/ (rs  0.331, p  0.036, n  43) and students in English Studies at a Spanish university, performed a negatively correlated with the fit index of English /ɪ/ to Spanish /i/ perceptual assimilation task evaluating the perceived similarity (rs  −0.333, p  0.029, n  43). Accurate /iː/-/ɪ/ discrimination between English /iː, ɪ, ε, ɜː/ and the perceptually closest Spanish was moderately correlated with the fit index of English /iː/ to vowels. English /ε, ɜː/ were included for comparison purposes. Spanish /i/ (rs  0.364, p  0.016, n  43). Similarly, /ε/-/ɜː/ Previous studies had found that English /iː/ and /ε/ have a clear discrimination was related to a greater assimilation percentage of counterpart in the L1, Spanish /i/ and /e/, respectively, while English /ε/ to Spanish /e/ (rs  0.335, p  0.028, n  43). These English /ɪ/ and /ɜː/ are less consistently mapped onto an L1 vowel results show that subjects who identified English /iː/ more (see Table 1 above). The goal of the current study was thus to successfully tended to assimilate English /ɪ/ to Spanish /e/ contrast the perceived similarity judgements obtained by a group rather than to Spanish /i/, and that subjects who were more of Spanish speakers with these same speakers’ ability to identify successful at discriminating English /iː/-/ɪ/ and English /ε/-/ɜː/ the English vowels, discriminate each vowel pair (/iː, ɪ/, /ε, ɜː/) tended to assimilate English /iː/ to Spanish /i/ and English /ε/ to and produce each vowel accurately. In addition, this study Spanish /e/, respectively, more consistently. However, there was investigated if the individual variation often observed in no correlation involving the perceived similarity results obtained perceived similarity was related to the variability found in L2 for /ɪ/ and /ɜː/ and the identification scores obtained for these performance. vowels. No correlations were found between the cross-linguistic The results of the perceptual assimilation task confirmed similarity measures and the production measures. On the other previous findings (e.g. Cebrian, 2019) and showed that the hand, moderate correlations were found between the results for English vowels differed in the degree to which they assimilated different vowels in the same task. Identification of English /ε/ was perceptually to Spanish categories. English /iː/ and /ε/ were correlated with the identification of /ɜː/ (rs 0 .323, p  0.035, n  strongly assimilated to Spanish /i/ and /e/ (93 and 99%, 43) and of /ɪ/ (rs  0.329, p  0.031, n  43). Within production, a respectively). By contrast, English /ɪ/ and /ɜː/ patterned as greater Euclidean distance between /iː/ and /ɪ/ was negatively more dissimilar to an L1 vowel, with /ɜː/ barely reaching a correlated with the degree of difference between an L2 speaker’s categorization threshold of 70% (72%) and /ɪ/ being slightly production and the native English mean in the case of /iː/ below that threshold (67%). Both /ɪ/ and /ɜː/ obtained low production (rs  −0.370, p  0.022, n  38) and /ɪ/ production goodness of fit scores, and consequently comparatively low fit (rs  −0.639, p  0.000, n  38). This means that the more indices (4/7 and 3.1/7, respectively; see Table 2 above for all differently participants produced /ɪ/ and /iː/, the more their results). Based on their strong assimilation to Spanish /i/ and productions resembled the native speakers’ production. Those /e/, it was hypothesized that English /iː/ and /ε/ would be participants who produced /ɪ/ more accurately (that is, with perceived and produced by the L2 learners with similarly high values closer to those of native speakers’) also tended to levels of accuracy. Regarding the poorly assimilated vowels, produce /iː/ more accurately, though the correlation in this perception and production accuracy may depend on the case was marginal (rs  0.320, p  0.05, n  38). extent to which the L2 speakers have detected differences In conclusion, perceived similarity does not seem to determine between the L1 and the L2 phones to establish separate accuracy of L2 vowel perception or production at an individual categories for these vowels (Flege, 1995). This in turn may level. The difficulty with the /iː/-/ɪ/ contrast, in particular, does depend on the amount of authentic L2 input received as well not seem to follow from the pattern of assimilation of each of as the amount of L2 use (Flege and Bohn, 2021). Given that L2 these two vowels to L1 categories. Some possible explanations for learning took place in an instructional setting in the L1 this fact are discussed next. country, with overall limited amount of L2 exposure and use, it is likely that learners were at a stage when the most dissimilar vowels are still challenging. In terms of GENERAL DISCUSSION discrimination ability, the PAM-L2 predicted that for L2 contrasts classified as uncategorized-categorized with Previous studies have found that Catalan/Spanish learners of partial overlap (/iː/-/ɪ/) and category-goodness assimilation English, in addition to learners of other language backgrounds, (/ε/-/ɜː/), discrimination accuracy depends on the degree of have difficulty distinguishing the English /iː/-/ɪ/ contrast both cross-language assimilation overlap (Faris et al., 2018; Tyler, in perception and in production (e.g., Flege, 1991; Flege et al., 2021). Thus, the expectation was that /iː/-/ɪ/ would be better 1997; Escudero and Boersma, 2004; Cebrian, 2006; Cebrian, discriminated than /ε/-/ɜː/, given the greater overlap Frontiers in Communication | www.frontiersin.org 12 June 2021 | Volume 6 | Article 660917 Cebrian et al. When the Easy Becomes Difficult displayed by the latter (72% as Spanish /e/) than the former duration difference produced by NES and also by other (32% as Spanish /i/). learners of English of a similar background reported in the The results of the vowel identification test showed that English literature (Cebrian, 2007; Rallo Fabra and Romero, 2012). The /ε/ was the most accurately identified vowel, while /ɜː/ was the possible role of duration in the categorization of the English /iː/-/ɪ/ worst (90 and 47%, respectively). This result may follow from the contrast is further discussed below. assimilation PAT results. However, the results for /iː/ and /ɪ/ did One of the goals of this study was to assess if variability in L2 not follow the expectations, and, in fact, the latter was better production and perception was related to individual differences identified than the former (/ɪ/: 72%, /iː/: 58%). This outcome is in in cross-linguistic perceived similarity. The acquisition of the line with previous research on the effect of perceptual training English /iː/-/ɪ/ contrast by Spanish/Catalan learners of English that found that Catalan/Spanish learners of English identified /ɪ/ provides a good ground to explore this issue as previous studies more successfully than /iː/ and /ɜː/ (Cebrian and Carlet, 2014; show inconsistent patterns of assimilation of English /ɪ/ to L1 Carlet and Cebrian, 2019; Fouz-González and Mompean, 2020) vowels. Recall that English /ɪ/ is generally assimilated to Spanish and that identification of /iː/ and /ɜː/ improved the most after /i/ and /e/ with varying degrees across studies. Assimilation of participants underwent perceptual training. The discrimination English /ɪ/ to Spanish /e/ ranges from 50% or less to 90% or more results also failed to support the predictions, as /ε/-/ɜː/ was more (see Table 1 above). The current study found that English /ɪ/ was accurately discriminated than /iː/-/ɪ/ (87% vs. 66%, respectively). identified with Spanish /e/ two thirds of the time, and with /i/ The discrimination results were replicated in production, as /ε/ one third of the time. An inspection of individual data showed and /ɜː/ were also more accurately produced than /iː/ and /ɪ/, that more than half the participants selected Spanish /e/ as the which tended to be produced as /iː/. In the case of /ε/, results were closest L1 vowel to English /ɪ/, while about a third selected consistent across tasks and in agreement with previous studies Spanish /i/ and the remaining were not consistent and selected involving Spanish L2 English speakers (Flege et al., 1997; Cebrian, both Spanish /i/ and /e/. Therefore, it was possible that these 2007). Further studies could investigate if this Spanish vowel differences in perceptual assimilation influenced the way learners would be undetectable by native English listeners when produced perceived and produced the English vowels. For instance, in a native English context (Flege, 1992). In fact, our preliminary participants who predominantly perceived English /ɪ/ as data from native English listeners yielded very high correct Spanish /e/ may be more successful at discriminating between identification scores for the learners’ /ε/ production (100%, /ɪ/ and /iː/ than those who assimilate both English /iː/ and /ɪ/ as with an average goodness rating of 5.8/7). The results for /ɜː/ Spanish /i/. However, the correlational analysis involving the showed a surprisingly high level of accuracy, in contrast with its perceptual similarity measures, the identification and poor identification results. Few studies have examined the discrimination results and the production results yielded very production of this vowel by Spanish/Catalan learners, but little evidence of a relationship between individuals’ cross- Carlet and Cebrian (2019) obtained a similar result with a linguistic perceived similarity and their L2 performance. Those similar population (second-year undergraduate students in an learners who assimilated English /ɪ/ to Spanish /e/ and who English Studies degree, as opposed to first-year undergraduates in judged English /iː/ to be closest to Spanish /i/, tended to be the current study). The relative success in the production of /ɜː/ better at identifying English /iː/ and discriminating between /iː/ may indicate that learners are capable of producing a vowel that is and /ɪ/. Still, no correlation involving the assimilation patterns different from other L2 (and L1) vowels, even if they cannot obtained for /ɪ/ and /ɜː/ and their identification or production successfully identify it. This difference may be methodologically accuracy were found. Therefore, the current study presents only a based, as identification involves the ability to distinguish the weak relationship between cross-linguistic perceived similarity target sound from other potentially conflicting sounds (e.g., other and L2 performance at an individual level and hence it does not response options present in the identification task), while provide evidence to the claim that learners’ accuracy in production involved the articulation of the sound in a high perception and production is related to individual variation in frequency word, which may make production more successful. the perceived similarity between L1 and L2 sounds (Flege and The idea that accurate perception precedes accurate production Bohn, 2021). Future research is necessary to examine if a in L2 (e.g., models like Flege (1995) SLM or Best (1995) PAM) is relationship can be found with other measures of similarity not always supported by the findings (e.g., Llisterri, 1995), and and of production and perception and with learners of more recent proposals suggest that L2 perception and production different levels of experience. may develop without the requirement that one modality precedes In summary, the perception and production of the English the other (Flege and Bohn, 2021, SLM-r). /iː/-/ɪ/ contrast by the L2 speakers in the current study does not The more striking results involve the vowel contrast that was follow from the perceived similarity relationships between the the focus of this study, the English /iː/-/ɪ/ contrast. In line with target and the native vowels, neither at a group level nor at an some previous studies, the results showed a high degree of individual level. This outcome is consistent with the results perceptual confusion between /iː/ and /ɪ/, which achieved only reported in previous studies involving Spanish and Catalan 66% correct discrimination, and were often misidentified as one learners of English showing comparatively poor perception another in the identification task. In addition, the two vowels were and production of both /iː/-/ɪ/, as discussed in Results and produced with a large amount of spectral overlap. In terms of Interim Discussions. In fact, similar results are reported for duration, L2 learners’ tense vowels were longer than the lax learners of other L1 backgrounds like Italian and Japanese vowels, but this difference in duration was smaller than the (e.g., Flege et al., 1998; Grenon et al., 2019). Frontiers in Communication | www.frontiersin.org 13 June 2021 | Volume 6 | Article 660917 Cebrian et al. When the Easy Becomes Difficult What is it then about the English /iː/-/ɪ/ contrast that makes it /ɪ/ and Spanish or Catalan /i/, their L2 category starts to drift difficult for L2 speakers to acquire? And why is it that a target vowel toward a more /ɪ/-like vowel. However, since at this stage both L2 like English /iː/ that is perceptually very close to an L1 vowel is vowels, /iː/-/ɪ/, share a single spectral category, possibly often found to be poorly perceived and produced? Flege (2018) distinguished temporally, the drift toward /ɪ/ affects the explains that cases of inaccurate L2 performance that have been perception and production of not only /ɪ/, but also /iː/. This attributed to variables such as a late starting age of learning or would result in a “deterioration” of /iː/, a vowel that by itself number of years of L2 use can be accounted for in terms of the should not be problematic given the high assimilation rates to an quality or quantity of the input, that is, whether learners are L1 vowel. In a similar vein, Major (1987) found that Brazilian exposed to authentic native-like input or to variable and often Portuguese speakers produced the similar English vowel /ε/ with accented input (see also Flege and Bohn, 2021). For example, increasing less accuracy as their production of the new vowel /æ/ Casillas (2015) reports that early Spanish-English bilinguals, who improved. In any event, the comparatively poor performance with had been exposed to English since age 4–6 and were dominant in English /iː/ shows that L2 vowels are not learned individually but as English as adults still differed from native English speakers in their part of a system of contrasting phones. This factor, together with use of spectral and temporal cues in their perception (but not the the effect of non-linguistic factors like the roles of orthography and production) of English /iː/-/ɪ/. A possible explanation offered is explicit instruction, may account for why cross-linguistic similarity exposure to Spanish-accented English. Morrison (2012), who also may not always make the right predictions, and why a target vowel found that Spanish speakers identified English /iː/ and /ɪ/ with that should in principle be “easy” to learn on the basis of its strong Spanish /i/ and /e/, respectively, argued that a perceptual perceptual assimilation to an L1 vowel, like English /iː/ for Spanish explanation for Spanish speakers’ difficulty with the /iː/-/ɪ/ speakers, is not accurately perceived and produced in the L2. contrast is not satisfactory and points to the effects of pronunciation instruction and orthography as possible causes. Indeed, orthography has been found to influence the IMPLICATIONS, LIMITATIONS AND pronunciation of L2 words (e.g., Morrison, 2009; Escudero and CONCLUSION Wanrooij, 2010). This influence may be particularly notable in foreign language learning contexts where L2 words are often first The observations made in the previous section have some encountered in writing. For example, the fact that the English lax pedagogical implications for the teaching and learning of this vowel /ɪ/ is typically spelt with the letter , which represents the vowel contrast. First of all, the contrast between English /iː/ and sound /i/ in many languages like Spanish, Catalan and Italian, may /ɪ/ should be set as a priority in L2 speech teaching given the result in the mispronunciation of this vowel (Morrison, 2008). contrast’s high functional load, and, consequently, the fact that English /ε/, on the other hand, is often spelt with the same letter as mispronunciation may contribute to loss of intelligibility and in Spanish, . The effect of orthography is likely enhanced by the reduction of comprehensibility [Brown, 1988; Levis, 2018; Munro, presence of many cognate words, particularly for speakers of 2021 (this volume); O’Brien, 2021 (this volume)]. Secondly, Romance languages (e.g., words containing such as cinema, instructors should avoid characterizing the English vowels /iː/ city, or minimum). Further, the tense vowel /iː/ is often represented and /ɪ/ as “long” and “short” respectively, as this may lead to an by a two-letter grapheme, e.g., in feet or in beat, which inaccurate simplification of the difference between them and may may suggest a longer duration. In addition, the effect of explicit make learners ignore vowel quality differences. Although the instruction that describes the contrast as merely a duration contrast duration difference is an important feature to illustrate, the focus (e.g., long /i/ vs. short /i/) may result in a misrepresentation of the should be on the qualitative difference. In that sense, the results of tense-lax contrast as a purely duration contrast (Flege et al., 1997; cross-linguistic perceived similarity tasks could be useful. For Wang and Munro, 1999). instance, given that English /ɪ/ is perceptually closer to Spanish As discussed in Introduction, L2 English speakers’ have been /e/ than to Spanish /i/, learners’ attention could be drawn to this L1 found to rely on duration to implement the English /iː/-/ɪ/ (e.g., vowel quality difference as a way of helping learners detect Escudero and Boersma, 2004; Cebrian, 2006; Kondaurova and differences between English /iː/ and /ɪ/. On a similar note, Francis, 2008). A combination of factors may account for this fact. cognate words could be used to illustrate differences in L1 and In addition to the issues described above, there are linguistic factors L2 pronunciation (e.g., Spanish “lista, video” and English “list, such as sensitivity to phonetic temporal distinctions in the L1, e.g., video”). Further, instruction should avoid exercises involving the as a result of stress or final obstruent voicing (Kondaurova and written form of words, at least at initial stages, and concentrate on Francis, 2008), desensitization to spectral contrasts not used in the auditory input as much as possible, in order to avoid the confusion L1 (Bohn, 1995), and statistical learning of the characteristics of that spelling might cause in the acquisition of this contrast. The use each vowel with the consequent detection that /iː/ tends to be of phonetic symbols is a possible option; Fouz-González and longer than /ɪ/ (Escudero and Boersma, 2004; Morrison, 2008). For Mompean (2020) report that both the use of keywords and example, Morrison suggests that learners go through different phonetic symbols in high variability phonetic training enhances stages in the process of learning the /iː/-/ɪ/ contrast which go identification accuracy of L2 vowels. See O’Brien (2021) (this from the inability to discern the two vowels, to establishing a volume) for an overview of approaches to and suggestions for duration contrast, to gradually detecting and incorporating spectral pronunciation teaching. Finally, note that, as Munro (2021) (this differences. Following this interpretation, we can speculate that volume) demonstrates, while there is some systematicity in the L2 when learners start to detect spectral differences between English pronunciation difficulties observed for a given L1 group, there is Frontiers in Communication | www.frontiersin.org 14 June 2021 | Volume 6 | Article 660917 Cebrian et al. When the Easy Becomes Difficult considerable variation in the pronunciation difficulties experienced readily perceived and produced successfully in terms of the L1 by different individuals. This fact underscores the importance of category (e.g., English /ε/ by Spanish speakers). Further, similarity individual pronunciation diagnosis and pronunciation practice predictions may differ for production and perception accuracy, as targeting individual difficulties. was the case of /ɜː/, a vowel that was very poorly perceived but The current study has some limitations that should be relatively accurately produced, which may be related to different acknowledged. The first limitation is the cross-sectional nature demands for perception and production tasks. The results for the of the current study. A single group of learners with a similar level /iː/-/ɪ/ contrast suggest that different factors are at play in the of proficiency was tested at a single point in time. Additional acquisition of this L2 contrast, including non-native overreliance groups differing in amount of experience or a longitudinal on acoustic cues (triggered by statistical learning or by approach would be necessary to fully evaluate the relationship metalinguistic information present in pronunciation between perceived similarity and L2 performance. In fact, L2 instruction), the effect of orthography, and the need to experience has been found to influence the way learners maintain an L2 contrast with a high functional load. These categorize the /iː/-/ɪ/ contrast (Aliaga-García and Mora, 2009; factors may override the effect of cross-linguistic similarity, as Ylinen et al., 2009; Grenon et al., 2019). For instance, Grenon the need to categorize the most dissimilar sound may bring with it et al. found that a subset of the inexperienced Japanese learners of the deterioration of the more similar sound, at least at some stages. English trained to attend to spectral differences between English These issues may be behind the reason why a presumably “easy” /iː/ and /ɪ/ managed to modify their original reliance on duration sound like English /iː/ may be more challenging than expected. and to attend to spectral cues, although their perception of the English contrast was not completely native-like. Another shortcoming involves the lack of L1 production data to DATA AVAILABILITY STATEMENT compare to the L2 data so as to have a measure of how differently the L1 and L2 vowels are produced, particularly Data from this study are available by request if required to when evaluating cases of near-identical L1-L2 vowels. An form part of a metaanalysis or similar form of research. additional limitation is related to the bilingual nature of the Requests to access the datasets should be directed to participants in this study, as, in addition to speaking Spanish, juli.cebrian@uab.cat. many participants spoke Catalan, with varying degrees of dominance (see Participants). Given the difference in vowel inventory between the two languages (five Spanish vowels vs. ETHICS STATEMENT seven Catalan vowels), it is difficult to determine, for instance, if the successful perception and production of English /ε/ is related Ethical review and approval was not required for the study on to the perceived similarity between this English vowel and Spanish human participants in accordance with the local legislation /e/ or Catalan /ε/4. Even if differences in Catalan/Spanish and institutional requirements. The patients/participants dominance were not found to affect similarity judgements (see provided their written informed consent to participate in Perceptual Assimilation Tasks Results), and the results of the PAT this study. in the current study closely replicated the results obtained by Cebrian (2019) with monolingual Spanish speakers, there is a potential effect of speaking more than one language on L2 AUTHOR CONTRIBUTIONS perception and production. Further, production data was mainly measured acoustically instead of by means of native JC contributed to conception and design of the study. JC, CG and speaker judgements, a measure that has been argued to be more NG collected the data, analyzed the results and performed the appropriate for measuring L2 production and accentedness statistical analysis. JC wrote the first draft of the manuscript. CG (Munro, 2008). Finally, the limited number of vowels examined, and NG wrote sections of the manuscript. two vowel contrasts, prevents the comparison of the current results with those for other vowel combinations, for example in the discrimination task. These issues are left for further research. FUNDING To conclude, the perception and production of L2 phones do not always follow from cross-linguistic similarity relations, This work was supported by research grants FFI2017-88016-P contrary to the predictions of most current L2 speech models. the Spanish Ministry of Economy and Competitiveness, and by In addition, no clear explanation has been found based on the research group Grant 2017SGR34 from the Catalan individual variation, as variability in L2 to L1 mapping does Government. not seem to be related to the degree of L2 perception or production accuracy, at least with the measures and the level of proficiency evaluated in this study. Still, some sounds may be SUPPLEMENTARY MATERIAL The Supplementary Material for this article can be found online at: 4 See Cebrian (2021) for a study on the perceived similarity between a near- https://www.frontiersin.org/articles/10.3389/fcomm.2021.660917/ complete set of Catalan and SSBE vowels. full#supplementary-material Frontiers in Communication | www.frontiersin.org 15 June 2021 | Volume 6 | Article 660917 Cebrian et al. When the Easy Becomes Difficult REFERENCES Casillas, J. (2015). Production and Perception of the /i/-/I/ Vowel Contrast: The Case of L2-Dominant Early Learners of English. Phonetica 72, 182–205. doi:10. 1159/000431101 Adank, P., Smits, R., and van Hout, R. (2004). A Comparison of Vowel Cebrian, J. (2006). Experience and the Use of Non-Native Duration in L2 Vowel Normalization Procedures for Language Variation Research. J. Acoust. Soc. Categorization. J. Phon. 34, 372–387. doi:10.1016/j.wocn.2005.08.003 Am. 116, 3099–3107. doi:10.1121/1.1795335 Cebrian, J. (2007). “Old Sounds in New Contrasts: L2 Production of the English Aliaga-García, C. (2011). “Measuring Perceptual Cue Weighting after Training: A Tense-Tax Vowel Distinction,” in Proceedings of the 16th International Comparison of Auditory vs Articulatory Training Methods,” in Achievements Congress of Phonetic Sciences. Editors J. Trouvain and W. J. Barry and Perspectives in the Acquisition of Second Language Speech: New Sounds (Saarbrucken, Germany: University of Saarland), 1637–1640. 2010. Editors M. Wrembel, M. Kul, and K. Dziubalska-Kołaczyk (Frankfurt am Cebrian, J. (2019). Perceptual Assimilation of British English Vowels to Spanish Main, Germany: Peter Lang), 15–28. Monophthongs and Diphthongs. J. Acoust. Soc. Am. 145, EL52–EL58. doi:10. Aliaga-García, C., and Mora, J. C. (2009). “Assessing the Effects of Phonetic 1121/1.5087645 Training on L2 Sound Perception and Production,” in Recent Research in Cebrian, J. (2021). Perception of English and Catalan Vowels by English and Second Language Phonetics/Phonology: Perception and Production. Editors Catalan Listeners: A Study of Reciprocal Cross-Linguistic Similarity. J. Acoust. M. A. Watkins, A. S. Rauber, and B. O. Baptista (Newcastle upon Tyne, Soc. Am. 149 (4), 2671–2685. doi:10.1121/10.0004257 United Kingdom: Cambridge Scholars Publishing), 2–31. Cebrian, J., and Carlet, A. (2014). Second-Language Learners’ Identification of Aoyama, K., Flege, J. E., Guion, S. G., Akahane-Yamada, R., and Yamada, T. (2004). Target-Language Phonemes: A Short-Term Phonetic Training Study. Can. Perceived Phonetic Dissimilarity and L2 Speech Learning: The Case of Japanese /r/ Mod. Lang. Rev. 70 (4), 474–499. doi:10.3138/cmlr.2318 and English /l/ and /r/. J. Phon. 32 (2), 233–250. doi:10.1016/S0095-4470(03)00036-6 Cerviño, E., and Mora, J. C. (2009). Duration as a Phonetic Cue in the Baigorri, M., Campanelli, L., and Levy, E. S. (2019). Perception of American- Categorization of /i:/-/ɪ/ and /s/-/z/ by Spanish/Catalan Learners of English. English Vowels by Early and Late Spanish-English Bilinguals. Lang. Speech. 62 J. Acoust. Soc. Am. 125, 2764. doi:10.1121/1.4784683 (4), 681–700. doi:10.1177/0023830918806933 Derwing, T. M., and Munro, M. J. (2015). A Prospectus for Pronunciation Research Best, C. T. (1995). “A Direct Realist View of Cross-Language Speech Perception,” in in the 21st Century. A Point of View. J. Second Lang. Pronunciation 1 (1), Speech Perception and Linguistic Experience: Issues in Cross Language Research. 11–42. doi:10.1075/jslp.1.1.01mun Editor W. Strange (Timonium, MD: York Press), 171–204. Escudero, P. (2005). Linguistic Perception and Second Language Acquisition: Best, C. T., and Strange, W. (1992). Effects of Phonological and Phonetic Factors on Explaining the Attainment of Optimal Phonological Categorization. PhD Cross-Language Perception of Approximants. J. Phon. 20 (3), 305–330. doi:10. Thesis. Amsterdam (Netherlands): Netherlands Graduate School of Linguistics. 1016/S0095-4470(19)30637-0 Escudero, P., and Boersma, P. (2004). Bridging the Gap Between L2 Speech Best, C. T., and Tyler, M. D. (2007). “Nonnative and Second-Language Speech Perception Research and Phonological Theory. Stud. Second Lang. Acquis. Perception,” in Honor of James Emil Flege. Editors O. S. Bohn and M. J. Munro 26 (4), 551–585. doi:10.1017/S0272263104040021 (Amsterdam, Netherlands: John Benjamins), 13–34. doi:10.1075/lllt.17.07bes Escudero, P., and Chládková, K. (2010). Spanish Listeners’ Perception of American Birdsong, D., Gertken, L. B., and Amengual, M. (2012). Bilingual Language Profile: and Southern British English Vowels. J. Acoust. Soc. Am. 128, EL254–EL260. An Easy-to-Use Instrument to Assess Bilingualism. Austin, TX: COERLL, doi:10.1121/1.3488794 University of Texas at Austin. Escudero, P., and Wanrooij, K. (2010). The Effect of L1 Orthography on Non- Boersma, P., and Weenink, D. (2018). Praat: Doing Phonetics by Computer Native Vowel Perception. Lang. Speech. 53 (3), 343–365. doi:10.1177/ [Computer Program]. version 6.0.43. Available at: http://www.praat.org/ 0023830910371447 (Accessed September 8, 2018). Faris, M. M., Best, C. T., and Tyler, M. D. (2018). Discrimination of Uncategorised Bohn, O.-S. (1995). “Cross-Language Speech Perception in Adults: First Language Transfer Non-native Vowel Contrasts is Modulated by Perceived Overlap with Native Doesn’t Tell it All,” in Speech Perception and Linguistic Experience: Issues in Cross Phonological Categories. J. Phon. 70, 1–19. doi:10.1016/j.wocn.2018.05.003 Language Research. Editor W. Strange (Timonium, MD: York Press), 279–304. Flege, J. E. (1991). The Interlingual Identification of Spanish and English Vowels: Bohn, O.-S., and Flege, J. E. (1992). The Production of New and Similar Vowels by Orthographic Evidence. Q. J. Exp. Psychol. A. 43 (3), 701–731. doi:10.1080/ Adult German Learners of English. Stud. Second Lang. Acquis. 14 (02), 131–158. 14640749108400993 doi:10.1017/S0272263100010792 Flege, J. E. (1992). “Speech Learning in a Second Language,” in Phonological Bohn, O.-S., and Steinlen, A. K. (2003). “Consonantal Context Affects Cross- Development: Models, Research, Implications. Editors C. A. Ferguson, L. Menn, Language Perception of Vowels,” in Proceedings of the XV International and C. Stoel-Gammon (Timonium, MD: York Press), 565–604. Congress of Phonetic Sciences. Editors D. Recasens, M. J. Solé, and Flege, J. E. (1995). “Second Language Speech Learning: Theory, Findings and J. Romero (Barcelona, Spain: Causal Productions), 2289–2292. Problems,” in Speech Perception and Linguistic Experience: Issues in Cross- Bradlow, A. R., Pisoni, D. B., Akahane-Yamada, R., and Tohkura, Y. i. (1997). Training Language Research. Editor W. Strange (Baltimore, MD: York Press), 233–277. Japanese Listeners to Identify English /r/ and /l/: IV. Some Effects of Perceptual Learning Flege, J. E. (2003). “Assessing Constraints on Second-Language Segmental on Speech Production. J. Acoust. Soc. Am. 101, 2299–2310. doi:10.1121/1.418276 Production and Perception,” in Phonetics and Phonology in Language Brown, A. (1988). Functional Load and the Teaching of Pronunciation. TESOL Q. Comprehension and Production. Editors A. Meyer and N. Schiller (Berlin, 22 (4), 593–606. doi:10.2307/3587258 Germany: Mouton de Gruyter), 319–355. Bundgaard-Nielsen, R. L., Best, C. T., and Tyler, M. D. (2011). Vocabulary Size Flege, J. E. (2018). It’s Input That Matters Most, Not Age. Biling. Lang. Cogn. 21 (5), Matters: The Assimilation of Second-Language Australian English Vowels to 919–920. doi:10.1017/S136672891800010X First-Language Japanese Vowel Categories. Appl. Psycholinguist. 32, 51–67. Flege, J. E., Frieda, E. M., Walley, A. C., and Fandazza, L. A. (1998). Lexical Factors doi:10.1017/S0142716410000287 and Segmental Accuracy in Second Language Speech Production. Stud. Second Busà, M. G. (1992). “On the Production of English Vowels by Italian Speakers with Lang. Acquis. 20 (2), 155–187. doi:10.1017/S0272263198002034 Different Degrees of Accent,” in New Sounds 92, Proceedings of the 1992 Flege, J. E., Bohn, O.-S., and Jang, S. (1997). Effects of Experience on Non-Native Amsterdam Symposium on the Acquisition of Second-Language Speech. Editors Speakers’ Production and Perception of English Vowels. J. Phon. 25 (4), J. Leather and J. Allan (Amsterdam, Netherlands: University of Amsterdam), 47–63. 437–470. doi:10.1006/jpho.1997.0052 Carlet, A., and Cebrian, J. (2019). “Assessing the Effect of Perceptual Training on L2 Vowel Flege, J. E., and Bohn, O.-S. (1989). The Perception of English Vowels by Native Identification, Generalization and Long-Term Effects,” in A Sound Approach to Spanish Speakers. J. Acoust. Soc. Am. 85 (1), S85. doi:10.1121/1.2027177 Language Matters. In Honor of Ocke-Schwen Bohn. Editors A. M. Nyvad, M. Hejná, Flege, J. E., and Bohn, O.-S. (2021). “The Revised Speech Learning Model (SLM-R),” A. Højen, A. B. Jespersen, and M. H. Sørensen (Aarhus, Denmark: Deparment of in Second Language Speech Learning: Theoretical and Empirical Progress. Editor English, School of Communication and Culture, Aarhus University), 91–119. R. Wayland (Cambridge, United Kingdom: Cambridge University Press), 3–83. Carlet, A., and Kivistö-de Souza, H. (2018). Improving L2 Pronunciation Inside Flege, J. E., Schirru, C., and MacKay, I. R. A. (2003). Interaction Between the Native and Outside the Classroom. Ilha do Desterro. 71 (3), 99–124. doi:10.5007/2175- and Second Language Phonetic Subsystems. Speech Commun. 40 (4), 467–491. 8026.2018v71n3p99 doi:10.1016/S0167-6393(02)00128-0 Frontiers in Communication | www.frontiersin.org 16 June 2021 | Volume 6 | Article 660917 Cebrian et al. When the Easy Becomes Difficult Fouz-González, J., and Mompean, J. A. (2020). Exploring the Potential of Phonetic Morrison, G. S. (2009). L1-Spanish Speakers’ Acquisition of the English /i/-/ɪ/ Symbols and Keywords as Labels for Perceptual Training. Stud. Second Lang. Contrast II: Perception of Vowel Inherent Spectral Change1. Lang. Speech. 52 Acquis. 1–32. doi:10.1017/S0272263120000455 (4), 437–462. doi:10.1177/0023830909336583 Fullana-Rivera, N., and MacKay, I. R. (2003). “Production of English Sounds by Morrison, G. S. (2012). Perception of Natural Vowels by Monolingual Canadian- EFL Learners: The Case of /i/ and /ɪ/,” in Proceedings of the 15th International English, Mexican-Spanish, and Peninsular-Spanish Listeners. Can. Acoust. 40 Congress of Phonetic Sciences, ICPhS. Editors M. J. Solé, D. Recasens, and (4), 29–40. doi:10.1121/1.3587895 J. Romero (Barcelona, Spain: Universitat Autònoma de Barcelona), 1525–1528. Munro, M. J. (2008). “7. Foreign Accent and Speech Intelligibility,” in Phonology Gilichinskaya, Y. D., and Strange, W. (2010). Perceptual Assimilation of American and Second Language Acquisition. Editors Hansen Edwards, J. G., and Zampini, English Vowels by Inexperienced Russian Listeners. J. Acoust. Soc. Am. 128, M. L. (Burnaby, Canada: Simon Fraser University), 193–218. EL80–EL85. doi:10.1121/1.3462988 Munro, M. J. (Forthcoming 2021). On the Difficulty of Defining “Difficult” in Second- Grenon, I., Kubota, M., and Sheppard, C. (2019). The Creation of a New Vowel Language Vowel Acquisition. Front. Commun. doi:10.3389/fcomm.2021.639398 Category by Adult Learners After Adaptive Phonetic Training. J. Phon. 72, Munro, M. J., Flege, J. E., and MacKay, I. R. A. (1996). The Effects of Age of Second 17–34. doi:10.1016/j.wocn.2018.10.005 Language Learning on the Production of English Vowels. Appl. Psycholinguist. Guion, S. G., Flege, J. E., Akahane-Yamada, R., and Pruitt, J. C. (2000). An 17, 313–334. doi:10.1017/s0142716400007967 Investigation of Current Models of Second Language Speech Perception: The O’Brien, M. G. (2021). Ease and Difficulty in L2 Pronunciation Teaching: A Mini- Case of Japanese Adults’ Perception of English Consonants. J. Acoust. Soc. Am. Review. Front. Commun. 5, 1–7. doi:10.3389/fcomm.2020.626985 107 (5), 2711–2724. doi:10.1121/1.428657 Piske, T., MacKay, I. R. A., and Flege, J. E. (2001). Factors Affecting Degree of Foreign Hacquard, V., Walter, M. A., and Marantz, A. (2007). The Effects of Inventory on Accent in an L2: A Review. J. Phon. 29 (2), 191–215. doi:10.1006/jpho.2001.0134 Vowel Perception in French and Spanish: An MEG Study. Brain Lang. 100 (3), Rallo Fabra, L., and Romero, J. (2012). Native Catalan Learners’ Perception and 295–300. doi:10.1016/j.bandl.2006.04.009 Production of English Vowels. J. Phon. 40, 491–508. doi:10.1016/j.wocn.2012.01.001 Higgins, J. (2018). Minimal Pairs for English RP: Lists by John Higgins. Available Recasens, D., and Espinosa, A. (2009). Dispersion and Variability in Catalan Five and at: http://minimal.marlodge.net/minimal.html (Accessed January 29, 2021). Six Peripheral Vowel Systems. Speech Commun. 51, 240–258. doi:10.1016/j. Hillenbrand, J. M., Clark, M. J., and Houde, R. A. (2000). Some Effects of Duration on specom.2008.09.002 Vowel Recognition. J. Acoust. Soc. Am. 108 (6), 3013–3022. doi:10.1121/1.1323463 Rochet, B. (1995). “Perception and Production of Second-Language Speech Sounds Højen, A., and Flege, J. E. (2006). Early Learners’ Discrimination of Second-Language by Adults,” in Speech Perception and Linguistic Experience: Issues in Cross Vowels. J. Acoust. Soc. Am. 119 (5), 3072–3084. doi:10.1121/1.2184289 Language Research. Editor W. Strange (Timonium, MD: York Press), 379–410. Holt, L. L., and Lotto, A. J. (2006). Cue Weighting in Auditory Categorization: Strange, W. (1995). Speech Perception and Linguistic Experience: Issues in Cross Implications for First and Second Language Acquisition. J. Acoust. Soc. Am. 119 Language Research. Timonium, MD: York Press. (5), 3059–3071. doi:10.1121/1.2188377 Strange, W., Hisagi, M., Akahane-Yamada, R., and Kubo, R. (2011). Cross- IBM Corp (2017). IBM SPSS Statistics for Windows. Version 25.0. Armonk, NY: Language Perceptual Similarity Predicts Categorial Discrimination of IBM Corp. American Vowels by Naïve Japanese Listeners. J. Acoust. Soc. Am. 130, Jun, S.-A., and Cowie, I. (1994). Interference for ‘New’ and ‘Similar’ Vowels in EL226–EL231. doi:10.1121/1.3630221 Korean Speakers of English. Ohio State Univ. Working Papers. 43, 117–130. Strange, W., Levy, E. S., and Law, F. F. (2009). Cross-Language Categorization of Kivistö-de Souza, H., Carlet, A., Julkowska, I., and Rato, A. (2017). Vowel French and German Vowels by Naïve American Listeners. J. Acoust. Soc. Am. Inventory Size Matters: Assessing Cue-Weighting in L2 Vowel Perception. 126 (3), 1461–1476. doi:10.1121/1.3179666 Ilha do Desterro. 70 (3), 33–46. doi:10.5007/2175-8026.2017v70n3p33 Thomas, E., and Kendall, T. (2007). NORM: The Vowel Normalization and Kondaurova, M. V., and Francis, A. L. (2008). The Relationship Between Native Plotting Suite. Available at: http://ncslaap.lib.ncsu.edu/tools/norm/ (Accessed Allophonic Experience With Vowel Duration and Perception of the English January 29, 2021). Tense/Lax Vowel Contrast by Spanish and Russian Listeners. J. Acoust. Soc. Am. Trubetzkoy, N. (1939). Principles of Phonology. Berkeley, CA: University of 124 (6), 3959–3971. doi:10.1121/1.2999341 California Press. Kuhl, P. K., and Iverson, P. (1995). “Linguistic Experience and the Perceptual Magnet Tyler, M. D. (2021). “Phonetic and Phonological Influences on the Discrimination Effect,” in Speech Perception and Linguistic Experience: Theoretical and of Non-Native Phones,” in Second Language Speech Learning: Theoretical and Methodological Issues. Editor W. Strange (Timonium, MD: York Press), 121–154. Empirical Progress. Editor R. Wayland (Cambridge, United Kingdom: Lado, R. (1957). Linguistics Across Cultures: Applied Linguistics for Language Cambridge University Press), 157–174. doi:10.1017/9781108886901.005 Teachers. Ann Arbor, MI: University of Michigan Press. Tyler, M. D., Best, C. T., Faber, A., and Levitt, A. G. (2014). Perceptual Assimilation Levis, J. M. (2018). Intelligibility, Oral Communication, and the Teaching of and Discrimination of Non-Native Vowel Contrasts. Phonetica 71, 4–21. doi:10. Pronunciation. Cambridge, United Kingdom: Cambridge University Press. 1159/000356237 Levy, E. S. (2009b). Language Experience and Consonantal Context Effects on Wang, X., and Munro, M. J. (1999). “The Perception of English Tense-Lax Vowel Perceptual Assimilation of French Vowels by American-English Learners of Pairs by Native Mandarin Speakers: The Effect of Training on Attention to French. J. Acoust. Soc. America. 125, 1138–1152. doi:10.1121/1.3050256 Temporal and Spectral Cues,” in Proceedings of the 14th International Congress Levy, E. S. (2009a). On the Assimilation-Discrimination Relationship in American of Phonetic Sciences. Editors J. Ohala, Y. Hasegawa, M. Ohala, D. Granville, and English Adults’ French Vowel Learning. J. Acoust. Soc. Am. 126 (5), 2670–2682. A. Baily (San Francisco, CA: University of California, Berkeley), 125–129. doi:10.1121/1.3224715 Ylinen, S., Uther, M., Latvala, A., Vepsäläinen, S., Iverson, P., Akahane-Yamada, R., Llisterri, J. (1995). “Relationships between Speech Production and Speech et al. (2009). Training the Brain to Weight Speech Cues Differently: A Study of Perception in a Second Language,” in Proceedings of the 13th International Finnish Second-Language Users of English. J. Cogn. Neurosci. 22 (6), Congress of Phonetic Sciences. Editors K. Elenius and P. Branderud 1319–1332. doi:10.1162/jocn.2009.21272 (Stockholm, Sweden: KTH and Stockholm University), 92–99. Lobanov, B. M. (1971). Classification of Russian Vowels Spoken by Different Conflict of Interest: The authors declare that the research was conducted in the Speakers. J. Acoust. Soc. Am. 49, 606–608. doi:10.1121/1.1912396 absence of any commercial or financial relationships that could be construed as a Major, R. C. (1987). English Voiceless Stop Production by Speakers of Brazilian potential conflict of interest. Portuguese. J. Phon. 15 (2), 197–202. doi:10.1016/S0095-4470(19)30560-1 Mora, J. C., and Fullana, N. (2007). “Production and Perception of English/iː/-/ɪ/ and /æ/-/ Copyright © 2021 Cebrian, Gorba and Gavaldà. This is an open-access article ʌ/ in a Formal Setting: Investigating the Effects of Experience and Starting Age,” in distributed under the terms of the Creative Commons Attribution License (CC BY). Proceedings of the 16th International Congress of Phonetic Sciences 3. Editors The use, distribution or reproduction in other forums is permitted, provided the J. Trouvain and W. Barry (Saarbrücken, Germany: University of Saarland), 1613–1616. original author(s) and the copyright owner(s) are credited and that the original Morrison, G. S. (2008). L1-Spanish Speakers’ Acquisition of the English /i /-/ɪ/ publication in this journal is cited, in accordance with accepted academic practice. Contrast: Duration-Based Perception is Not the Initial Developmental Stage. No use, distribution or reproduction is permitted which does not comply with Lang. Speech. 51 (4), 285–315. doi:10.1177/0023830908099067 these terms. Frontiers in Communication | www.frontiersin.org 17 June 2021 | Volume 6 | Article 660917