Acquiring native-like proficiency in a second language (L2) is difficult to achieve after the critical period (Lenneberg 1967). Acquisition of L2 intonation and prosody has been assumed to be the last stage of L2 acquisition and is one of the least explored areas of intonation research. Languages differ in their prosody, especially in the way they mark prominence. Languages like English and Spanish mark word prominence by pitch accent realized on the stressed syllable of a word (most often a content word). Languages like Korean, which does not have lexical or postlexical stress, mark word prominence by forming the word into one prosodic unit called the Accentual Phrase (AP) or by locating the word at the beginning of an AP (Jun 1993). In Jun’s prosodic typology (Jun 2005b, 2014b), the former type belongs to a head-prominence language, and the latter type belongs to an edge-prominence language.
In addition to this prosodic difference, Spanish and Korean differ in the division of lexicon. Spanish distinguishes between a content word and a function word, but Korean does not have function words. Instead, all words in Korean are content words, consisting of a lexical item plus a case marker or postpositions, which do not form a separate AP, but always form an AP together with the preceding lexical item.
Given these prosodic and morphological difference between the two languages, the present study is divided into four fundamental components. First, I posit that Korean Accentual Phrase (AP) initial tone-segment mapping will be applied to Spanish, in that the High tone will be applied when the phrase-initial segment is aspirated (/ph, th, kh, tʃh/), tense (/p*, t*, k*, s*, tʃ*/) or /h/ or /s/; otherwise, Low tone will be applied (Jun 1993, 1998, 2000, 2005). Secondly, I hypothesize that nuclear and pre-nuclear Pitch Accents in Spanish will be realized differently since Korean does not have this distinction. Thirdly, I posit that Korean learners will tend to produce an AP-like tonal unit over a function word as well as a content word. Lastly, I hypothesize that the boundary tones in Spanish in information-seeking yes-no questions and wh-questions, invitation yes-no questions and wh-questions will be realized as HH% or LH% due to L1 influence.
Twenty-one L2 participants representing three proficiency levels (seven beginners, seven intermediates, and seven advanced learners) recorded themselves reading words, phrases, and sentences in Spanish in as naturally as possible. Native Mexican and Peninsular Spanish speakers (three participants each) also participated as a control group.
Praat was used to analyze the intonation, and the pitch contour was labeled following the Spanish ToBI (Tones and Break Indices) (Beckman et al. 2002, Prieto & Roseano 2010) and Korean ToBI (Jun 2000, 2005) transcription conventions, which are based on the Autosegmental-Metrical model of intonational phonology (Pierrehumbert 1980, Beckman & Pierrehumbert 1986, Ladd 1996/2008) of each language.
The results showed that the Korean Accentual Phrase (AP) initial tone-segment mapping had a stronger effect on beginner and intermediate groups than advanced learners. Also, L1 Korean speakers had difficulties in marking prominence according to Spanish conventions and had a strong tendency to produce an “L H” (rising) tonal pattern over a word, i.e., the AP tone pattern, regardless of the location of stress. Stress-based pitch accent in Spanish is difficult to acquire for speakers of L1 Korean, which has no word-prosody. Korean AP tones, which cue word prominence via edge marking, are negatively transferred in producing a word and pitch accent in Spanish. Also, in Seoul Korean, as each word forms one Accentual Phrase (AP), L2 learners produced AP-like tonal units especially on monosyllabic function words in Spanish. It is interesting that the majority of the intermediate and advanced Native Korean Speakers (NKS) produced L% in information-seeking wh-questions and in invitation wh-questions, which is not what I had predicted based on previous research. This may be due to the English intonation of information-seeking wh-questions and invitation wh-questions, which use a falling tone. Future studies are needed to confirm these initial findings of the influence of English on L1 Koreans’ production. Additionally, according to the results, beginners tend to commit more errors, and I posit that this is because these types of sentences require conveying of meaning through intonation, beyond simple lexical items. According to these results, the four dimensions of Mennen (2015)’s L2 Intonation Learning theory (LILt)) were supported.
So far, no study has been carried out regarding the Acquisition of Spanish intonation by Native Korean Speakers based on the Autosegmental-Metrical model of intonational phonology, using Spanish ToBI and Korean ToBI transcription conventions. This study offers significant contributions to the field of L2 prosody acquisition of native Korean speakers learning Spanish, as my analysis is not limited to boundary tone realization, but extended to the inventory of structural phonological elements such as Pitch Accents in Spanish and Accentual Phrases in Korean. In general, L2 learners are not aware of the prosodic characteristics of their second language and tend to apply their native language’s prosodic features to their L2.