English speakers learning the Korean language pronounce words such as 막내
Consonantal sequences are defined as the case in which consonants are adjacent, and these occur within a syllable and also across a syllable boundary. For example, the consonantal sequences within a syllable include two kinds of cases, one is consonantal sequences that appear at an onset position such as ‘tr-, pl-’ in ‘tree, play’, and the other is sequences that appear at the end of a word such as ‘-nt, -mp’ in ‘tent’ and ‘camp’ in English. Unlike English, as is well known, consonantal sequences within a syllable are not possible in Korean (Heo, 1995; Kim 1996; Rhee, 2002). However, the other case in which consonantal sequences occur across a syllable boundary is common. In this case, the first consonant of the consonantal sequences is a final consonant of the preceding syllable, and the second consonant is an onset of the following syllable. For instance, in 안개
In this article, we investigate the latter case, i.e., consonantal sequences across a syllable boundary by roughly dividing them into two categories.1 The first group is one in which adjacent consonants do not affect each other, and thus no phonological changes take place. The second group is one in which consonantal assimilation takes place. This consonantal assimilation can also be divided into two cases. One case is when a consonant triggers assimilation in the place of articulation, while the other depends on the manner of articulation. Based on an analysis of the pronunciation types of consonantal sequences, we seek an effective way to teach the pronunciation of Korean consonantal sequences to foreigners.
1The case of obstruent-obstruent sequence is excluded in this article. In this case, the second consonant is realized as a tensed counterpart by the process of tensification, as shown in such examples as 국수 kwukswu [kuks’u], 식당 siktang [sikd’aŋ], 옆집 yephcip [yəpc’ip], and so forth.
As mentioned above, some languages have consonantal assimilation, others do not. In other words, some kinds of relationships between the two adjacent consonants are established in some languages, others are not. For instance, in Chinese and Vietnamese, each syllable is pronounced separately no matter what kinds of consonants are adjacent to each other (Lin, 2007; Chau, 2011). That is, a consonant cannot affect the adjacent consonant in terms of the place of articulation or the manner of articulation. In other words, in these languages, it is not a matter of which consonant is adjacent to which consonant. In most languages, however, the sound value of a consonant is changed by the effect of the adjacent consonant in terms of the place of articulation or the manner of articulation, or both.
In the table given in (1) below, Sino-Korean words with Chinese characters are presented with their corresponding Chinese pronunciation.
Pronunciation of Sino-Korean and Chinese with the same meaning
As is well known, Chinese, unlike Korean, allows [n] and [ŋ] only in the coda position of a syllable. So the case in which two consonants are in a row is not often found, i.e., only the consonants [n] and [ŋ] can be adjacent to other consonants. In Chinese, as can be seen above, no phonological changes between adjacent consonants are found. Each syllable is pronounced separately as it is without any changes. However, in Korean, the nasal-liquid sequence is pronounced as nasal-nasal as in 심리
Let us look at the examples of Korean and Vietnamese given in (2) below.
Pronunciation of Sino-Korean and Vietnamese with the same meaning
Unlike Chinese, in Vietnamese, most consonants can appear at the end of a word, if they are unreleased. However, as seen above, no phonological changes in adjacent consonants take place in any sequences. In the case of
In the case of nasal-liquid sequences such as
Place assimilation is defined as being what occurs when adjacent consonants pronounced in the different places of articulation are pronounced in the same place by the effect of a neighboring consonant. This includes Koreans’ pronounciation of 안개
In Japanese, consonantal assimilation has two different cases, one occurs between nasal /N/ marked as ‘ん’ and the following consonant, and the other occurs between obstruent /Q/ marked as ‘っ’ and the following consonant (Kwon, 2006; Vance, 2008).3 These two special consonants can appear only at the end of a syllable. They cannot appear word-initially. The nasal /N/ ‘ん’ can be pronounced as [m], [n], or [ŋ], the obstruent /Q/ ‘っ’ can be pronounced as [p], [t], or [k]. The pronunciation is identified by the place of articulation of the following consonant.
To be more concrete, let us consider the examples given below:
Pronunciation of 三 ( さん )
In Korean, the Chinese character 三(삼) ‘three’ is always pronounced as [sam] no matter what the following consonant is, whereas in Japanese, 三(さん) is pronounced as [sam], [san], and [saŋ] depending on the following consonant. /N/ ‘ん’ is realized as one of the allophones such as [m, n, ŋ] depending on the place of articulation of the following consonant. It is realized as a nasal consonant in the same place as the following consonant. That is, it would be pronounced as [m], [n], or [ŋ], respectively, when the place of articulation of the following consonant is labial, alveolar, or velar. This is supported by the following actual examples.
First of all, /N/ is realized as the labial nasal consonant [m] before a labial [m], [b], and [p].
In the words above, the final consonant of the first syllable in each word is a nasal /N / ‘ん’. Therefore, it is realized as a labial [m] because the following consonant is one of the members of the group of labial consonants such as [m], [b], and [p], respectively.
Secondly, /N/ is realized as the alveolar nasal [n] before an alveolar [s], [d], [z], [n], [t], and so on.
In all the examples above, the final consonant of the first syllable is a nasal /N/ ‘ん’, so it is realized as an alveolar nasal [n] assimilating the place of articulation of the consonants [s], [d], and [z].
Lastly, /N/ is realized as a velar nasal [ŋ] before a velar [k], [g].
The final consonant of the first syllable /N/ ‘ん’ is realized as a velar nasal [ŋ] assimilating the place of articulation of the following velar consonant such as [k], [g].
This causes Japanese native speakers difficulty in the pronunciation of consonantal sequences that have different places of articulation such as 안개
Let us now proceed to a discussion about 一(일) ‘one’. In Korean, the Chinese character 一(일) is pronounced as [il] at all times no matter what the following consonant is. However, in Japanese, as shown in the examples given below, 一(いっ) is pronounced as either [ip], [it], or [ik] depending on the following consonant.
Pronunciation of 一(いっ)
/Q/ ‘ っ ’ is realized as one of the allophones [p], [t], or [k] depending on the place of articulation of the following consonant. In other words, when the place of articulation of the following consonant is labial, alveolar, or velar, it would be pronounced as [p], [t], or [k], respectively. The pronunciation of /Q/ is identified by the place of articulation of the following consonant. As a result of this process, the consonantal sequence in Japanese is realized as geminates such as /-pp-, -tt-, -kk-/. Consonantal sequences which are composed of the different place of articulation are not well-formed.
To be more specific, let us consider the following examples:
First of all, /Q/ is realized as the labial consonant [p] before a labial [p].
In all the examples above, the final consonant of the first syllable is /Q/ ‘っ’, so it is realized as [p] assimilating the place of articulation of the following consonant.
Secondly, /Q/ is realized as the alveolar consonant [t] before an alveolar [t].
In the examples above, the final consonant of the first syllable ‘っ’ is realized as [t] assimilating the place of articulation of the following consonant.
Lastly, /Q/ is realized as the velar consonant [k] before a velar [k].
The final consonant of the first syllable is the obstruent ‘っ’. It is realized as [k] having the same place of articulation as the following consonant [k].
So far, we have seen that consonantal assimilation in terms of the place of articulation takes place in Japanese. Now, let us go on to a discussion of the case of assimilation in English.
Place assimilation in English can be found in the combination of the negation prefix ‘in-’ and root.4 The prefix ‘in-’ is pronounced differently depending on its phonological environment as shown below:
The pronunciation of negation prefix ‘in- ’ + root
The negation prefix ‘in-’ is pronounced as [in], [iŋ], [im] depending on the place of articulation of the following consonant. Especially, before the labial consonant, the change of spelling occurs as in ‘impossible and immoral’ (Lee, 2007; Kim, 1996). /N/ in the negation prefix ‘in-’ is pronounced as [n], [ŋ], or [m] assimilating the following consonant alveolar, velar, or labial, respectively.
So far, the relationship established by the place of articulation between adjacent consonants has been discussed. From this, we have seen that place assimilation takes place in Korean, but the change is not accepted in 표준발음법
2According to 어문규범 Emunkyupem ‘National Standards of Korean Language’ number 21, place assimilation is not permitted as in 감기 kamki [kamgi] not [kaŋgi], 옷감 oskam [otk’am] not [okk’am], 있고 issko [itk’o] not [ikk’o], 젖먹이 cecmeki [jənməki] not [jəmməki], 문법 munpep [munp’əp] not [mump’əp], 꽃밭 kkochpath [k’otp’at] not [k’opp’at]. 3Many of the Japanese examples in this article are taken from these two references. 4In English, the two consonants sharing the same place of articulation such as ‘-mp, -nt, rt, ŋk’ appear in a word-final position as seen in ‘camp, tent, cart, pin[ŋ]k’ For more detailed discussion about consonantal sequences in English, refer to An & Choi, 2006 and Heo & Kim, 2013, among others. These consonantal sequences, however, occur within a syllable forming a consonant cluster, and thus are not dealt with in this article.
Manner assimilation is defined as occurring when two consonants in a row pronounced by a different manner of articulation are realized by the same manner of articulation due to the neighboring consonant. Let us first consider the case of Korean.
Consonantal assimilation in terms of the manner of articulation is an obligatory phenomenon in Korean phonology, but only a few of languages have this process (Kim, 2008; Heo & Kim, 2013). In Korean, there are two kinds of consonantal assimilation in terms of the manner of articulation: namely, nasalization and lateralization. Nasalization takes place in words like 국민
First of all, the final consonants /k/, /t/, and /p/ are pronounced as [ŋ], [n], and [m], respectively, before a nasal /n/, /m/.6 Examples given below show the nasalization process of obstruent-nasal sequences.
Nasalization of obstruents in obstruent-nasal sequences
The interesting point here is that the preceding consonant is nasalized by the effect of the following consonant when a nasal consonant comes after an obstruent as in 막내
Secondly, the liquid /l/ after the final consonant /m/ or /ŋ/ is pronounced as a nasal [n]. Examples are presented below:
Nasalization of liquids in nasal-liquid sequences
However, when a nasal consonant comes after a liquid such as 설마
Finally, when the liquid /l/ comes after a following obstruent such as /k/, /p/, it is realized as its corresponding nasal retaining its place of articulation. That is, /k/ is realized as [ŋ], and /p/ is realized as [m]. Together with this, the liquid /l/ is pronounced as [n]. That is, the two consonants in sequence cause the change in the other by affecting each other.
Nasalization of obstruents and liquids in obstruent-liquid sequences
Yet, no phonological changes happen in cases where an obstruent such as /k/ or /p/ comes after a liquid /l/. This can be recognized in the pronunciation of such words as 딸기
This contrasts with the case of Vietnamese where the word like
To be more concrete, the consonantal sequences of obstruent-nasal (e.g., good news, nickname) and obstruent-liquid (e.g., pork ribs, upload) are possible in English. In Korean, however, an obstruent in the preceding consonant position is pronounced as a nasal which is pronounced in the same place of articulation, and a liquid is pronounced as a nasal [n]. That is, both consonants forming a consonantal sequence (e.g., 국민
To further clarify the principle of consonantal assimilation, we can compare the process of consonantal assimilation in terms of the manner of articulation in Korean and English. Let us first consider the condition of consonantal assimilation given below:
In English, the consonantal sequence ‘p-m’ and ‘k-n’ in pop music and nickname are pronounced as [pm] and [kn], not [mm] and [ŋn]. That is, the obstruents ‘p’ and ‘k’ do not undergo changes. When two neighboring consonants establish a relationship, the following consonant must not be weaker than the preceding one (Harris, 1994).
The strength of consonants is given below:
Consonant Hierarchy of Strength7
When the following consonant is weaker than the previous one, consonantal assimilation takes place to increase the strength of the following consonant. In the opposite case, no consonantal assimilation takes place. There are two ways to make the following consonant stronger than the preceding one. One is to increase the strength of the following consonant and the other is to decrease that of the preceding consonant. In the case of Korean, the strength of the preceding consonant is first weakened. In this case, regressive assimilation, whereby the following consonant affects the previous one, takes place. However, in the case in which weakening the preceding consonant is not possible, increasing the strength of the following consonant (i.e., progressive assimilation) takes place. Of importance in the adjustment of the consonantal strength is the fact that the place of articulation for the consonant in question is fixed. In other words, labials can be realized as labials, alveolars as alveolars, velars as velars. The strength hierarchy of Korean consonants with the place of articulation is presented in (16) below (Heo & Kim, 2006).
The strength hierarchy of consonants in Korean
In the two adjacent consonants, lateralization is also found in Korean. Nasal ‘n’ is pronounced as [l] before or after liquid ‘l’. Here are some relevant examples.
Lateralization of nasal /n/ in nasal-liquid and liquid-nasal sequences
Unlike nasalization, the order of ‘n’ and ‘l’ has no influence on whether the phenomenon of lateralization takes place or not. The nasal ‘n’ is always pronounced as the liquid [l] regardless of the order of ‘n’ and ‘l’. Lateralization is not found in Chinese and Vietnamese as the pronunciation is separated by the unit of syllables. In English, lateralization also does not take place in the same environment. Here are some relevant examples.
Pronunciation of nasal-liquid sequence in other languages
As mentioned earlier, consonantal assimilation in terms of the manner of articulation between the two consonants in Korean is not common in languages, unlike place assimilation. However, a similar process is observed in Hindi phonology. Let us briefly consider the process of Hindi nasalization.
The so-called consonant Sandhi in Hindi is very similar to the case of Korean consonantal assimilation (Choi, 2003). Here are some examples.
Examples of Hindi
In Hindi, consonantal sequences such as ‘k-m, t-n, t-m’ are not pronounced as they are, as is also the case in Korean. The first consonants must undergo nasalization. As a result, they are pronounced as [ŋm], [nn], [nm] just with the consonantal assimilation in terms of the manner of articulation. Consonantal assimilation in Hindi is recognized as the same phenomenon as occurs in Korean. Compare the pronunciation of the words in (19) above with that of words such as 국물
So far the pronunciation types of consonantal sequences have been surveyed. Korean is different from Vietnamese and Chinese that have no phonological change between two adjacent consonants. Korean is also different from English and Japanese that only have a connection with Korean in terms of assimilation depending on the place of articulation. Korean has the same phonological phenomenon as Hindi, however, in terms of consonantal assimilation sharing the manner of articulation between the two consonants in the sequence.
In the next section, we will consider how to teach pronunciation involving consonantal sequences to Korean language learners based on the discussion presented so far.
5For more detailed discussion about the phonological behavior of Korean consonants, including consonantal assimilation, refer to Kim (1996, 2003) and Heo (2004, 2007). 6To be more precise, the final consonant /k/ means the unreleased and simplified velar consonant, and thus includes ‘kk, k’, ks, lk’. The consonant /t/ includes, ‘ss, c, ch, t, h’ and /p/ includes ‘ph, lp, lph, ps’ for the same reason. 7This can also be explained by the notion of “sonority”. The sonority hierarchy is reversed, compared to the notion of “consonantal strength” based on Government Phonology (Harris, 1990; Harris & Kaye, 1990). That is, the order of sonority is as follows: obstruents< nasals
Eckman (1987) has accounted for relative degrees of difficulty in learning foreign languages. His Markedness Differential Hypothesis provided a useful explanation for determining the directionality of difficulty. According to Eckman, marked items in a language will be more difficult to learn than unmarked ones, and degrees of markedness will correspond to degrees of difficulty (Brown, 1994). Major and Faudree (1996) found that the phonological performance of Korean native speakers learning English reflected principles of marked universals.
Markedness is related to naturalness. That is, unnaturalness corresponds to markedness, while naturalness corresponds to unmarkedness. If something is marked, then its frequency is low. However, if something is unmarked, its frequency is high. For instance, voiced plosives are more marked than their voiceless counterparts. In this regard, it can be stated that syllable final neutralisation and consonantal assimilation in Korean are marked. Marked elements take more time and energy to acquire and cause more errors in the learning of foreign languages. However, the binary fission whereby something will ‘either be difficult or not’ to foreign language learners has somehow been supplemented by the Markedness Differential Hypothesis. Eckman attempted to provide an explanation of the difficulty of learning a language depending on the difference of markedness between L1 and L2.
Correlation between the markedness degree and the difficulty of learning L2
In learning target item β of L2, it is difficult to learn when the markedness degree of β is higher than α of L1 which is similar to L2. When the markedness degree is low, it is relatively easy to learn. For example, Korean has three different obstruents which are divided by the strength of the aspiration coming from one's lung, whereas English has two kinds of obstruents in terms of voicing. Both are fairly marked. However, according to the theory of Markedness Differential Hypothesis, because the markedness of Korean consonants is higher than that of English ones, Korean native speakers will experience more difficulties than English native speakers in learning the counterpart's consonants.
Consonantal assimilation in Korean is a very peculiar phenomenon which cannot easily be found in other languages, which means that its markedness degree is high. For this reason, Korean language learners are unfamiliar with this consonantal assimilation, and have some difficulty in its acquisition. It is important to be aware that the phonological change occurs in its manner, not in its place of articulation in teaching the pronunciation of consonantal sequences in Korean (Heo & Kim, 2006).
Let us now consider how to teach the pronunciation of Korean consonantal sequences. First, let the learners understand that Korean does not have place assimilation with the following words that have two different places of articulation. Here are some examples:
Consonantal sequences having different places of articulation
Learners who have languages with place assimilation as their mother tongue are prone to pronounce the words
The next step is to teach examples in which nasalization or lateralization takes place. The phonological environment of the two phonological processes is complicated and thus Korean language learners produce many errors and have difficulties in the process of acquisition. Let us first consider the case of nasalization. Words with nasalization can be divided into three categories depending on their phonological environment, as discussed previously. Teachers should teach the three categories one by one. If teachers present all the relevant examples at one time, students may get confused by the complex phonological environment.
Boonmalerd (2013) investigated the error rate found in the pronunciation of consonantal sequences of Thai students learning Korean. In the case of the ‘obstruent+nasal’ sequence, the error rate was 46.26%, and for the ‘nasal+liquid’ sequence it was 69.75%. However, the error rate of the ‘obstruent+liquid’ sequence went up to 85.1%. This means that the pronunciation of the sequence ‘obstruent+liquid’ is more difficult than other cases in the process of acquisition. Kim (2013) also presented the error rate found in the pronunciation of the consonantal sequences in an experiment with Chinese students learning Korean. Even if the exact percentage of errors is different from the result of Thai students, the overall characteristics are similar in both studies. These kinds of studies demonstrate that the order of acquisition of the consonantal sequences is ‘obstruent+nasal’, ‘nasal+liquid’, and ‘obstruent+liquid’. In other words, Korean language learners have the most difficulty in the acquisition of ‘obstruent + liquid’. This forces the order of teaching consonantal sequences of Korean involving nasalization to be ‘obstruent+nasal’, ‘nasal+liquid’, and ‘obstruent+liquid’.
Let us consider the first group, i.e., the sequence of ‘obstruent + nasal’. It would be helpful to let learners know that no phonological changes occur in the same phonological environment in English by comparing English words such as ‘nickname, good man, good night, map maker’. What is interesting is that nothing happens when a nasal is followed by an obstruent in Korean, as given in the examples below:
Sequence of ‘obstruent + nasal’
The second group is the sequence of ‘nasal+liquid’. It does not cause any phonological change in English either. Comparison with English words such as ‘kingly, some like’ which have no assimilation in the same phonological environment will help learners’ understanding. What is interesting is that no phonological processes take place when a liquid is followed by a nasal in Korean, as given in examples below:
Sequence of ‘nasal + liquid’
The last group is the sequence of ‘obstruent+liquid’. Unlike the former cases, there are no nasals in the two consonants in question, but both consonants are realized as nasal consonants, so that Korean language learners experience more difficulties. The comparison with English words such as ‘upload, pork ribs’ in which the obstruents do not undergo changes in the same phonological environment will help learners’ understanding. What is interesting is that no phonological processes take place when a liquid is followed by an obstruent in Korean. The words given below can be employed for teaching this aspect of Korean pronunciation that requires special attention.
Sequence of ‘obstruent + liquid’
Let us now consider how to teach the pronunciation of words involving lateralization. Teaching lateralization is relatively easier than teaching nasalization, since the sequence of ‘n’ and ‘l’ is changed into [ll] regardless of their order.8 Yet, it can be challenging because the pronunciation of ‘l’ in a row, i.e., [ll] is quite difficult for Korean language learners to pronounce. Therefore, repeated practice is necessary for them to get used to the pronunciation of the sequence ‘ll’. Here are some examples of lateralization.
The nature of lateralization in Korean can be acquired by comparing Korean examples with English examples in which lateralization does not take place as shown in ‘only, Henry, on-line, all night’ in the same phonological environment. Besides, students may easily grasp the concept of lateralization, if the pronunciation of each English word is pronounced as it would be if it were a Korean word.
8In Boonmalerd (2013), the error rate of lateralization made by Thai learners was 65.5% in ‘l+n’, 70.95% in ‘n+l’. It is assumed that the phonological environment is not complicated, but the pronunciation of liquids in a row ‘ll’ is not easy for Thai-speaking learners. A similar result is obtained in Kim (2013).
Each language pronounces consonantal sequences in a different way. Consonantal assimilation in terms of the manner of the articulation between the two adjacent consonants is abundant in Korean, unlike in other languages. For this reason, Korean language learners have a lot of difficulties in the acquisition of Korean pronunciation. To teach the pronunciation of consonantal sequences in Korean, it is vital to let the learners recognize the conditions where the assimilation takes place.
According to the theory of Markedness Differential Hypothesis, consonantal assimilation in Korean is an idiosyncratic property, and thus Korean language learners have some difficulty in the acquisition of the pronunciation of consonantal sequences involving consonantal assimilation. For this reason, a special strategy must be employed to teach the pronunciation of the consonantal sequences. In teaching the pronunciation of words involving nasalization, after dividing the examples into three categories, the phonological environment must be introduced one by one according to their level of difficulty. The teaching order of Korean consonantal sequences involving nasalization is ‘obstruent+nasal’, ‘nasal+liquid’, and then ‘obstruent+liquid’. In the case of words involving lateralization, the teaching focus must be placed on real practice of the pronunciation itself rather than on an understanding of the phonological environment.
Nasalization, particularly takes place not only within a word but also across a word boundary such as between objects and predicates or objects and adverbs. This is shown in such examples as 밥 먹어요