알타이어의 강조형용사는 어기(base)의 첫 CV의 중첩에 이어 {p, m, s, r}중의 하나의 자음을 연결하여 형성된다. 본 논문에서는 터키어와 몽고어에 일어나는 이 현상을 분석하여 그 근원을 밝히고자 한다. 우선 터키어의 강조형용사 중 어기가 모음으로 시작하는 경우를 ...

알타이어의 강조형용사는 어기(base)의 첫 CV의 중첩에 이어 {p, m, s, r}중의 하나의 자음을 연결하여 형성된다. 본 논문에서는 터키어와 몽고어에 일어나는 이 현상을 분석하여 그 근원을 밝히고자 한다. 우선 터키어의 강조형용사 중 어기가 모음으로 시작하는 경우를 어원적으로 분석해보면 [p]가 강조형용사 형성에서 삽입되는 것이 아니고 오히려 어원적 측면에서 보면 [p]가 어기의 일부분이었으나 알타이어에서는 어두의 [p]가 일반적으로 탈락하게 됨을 알 수 있다. 예를 들면 터키어 uzun 'long'은 *puzun에서 어두의 [p]가 탈락한 것인데 강조형용사에서는 이 원래의 어기중 첫 CV의 중복이 일어나 *pupuzun이 되고 uzun에서 처럼 어두의 [p]가 탈락함으로써 upuzun이 나오게 된 것이다. 공시적 관점에서 이러한 현상은 마치 (c)V의 중첩이후에 연결자음 [p]가 삽입된 것처럼 보이게 된다. 이러한 공시적 [p]삽입에 문법화가 일어나게 되면 원래 [p]로 시작하지 않은 어기도 문법화된 [p]삽입규칙의 적용을 받게 된다. 예: kara 'black' kap-kara 'pitch-black'. 다른 연결 자음도 비슷한 근원을 가질 수 있다고 본다. ciplak 'naked' --> sirilciplak 'entirely naked', karisik 'confused' --> karmakarisik 'completely confused'와 같은 또 다른 강조형용사의 예들은 연결자음 [m]과 [r]의 발생에 대하여 시사하는 바가 크다. 흥미롭게도 이 예들은 우리말의 '크나크다', '멀고멀다'등의 강조형용사와도 흡사하다. 이러한 강조형용사 구문에 중간모음탈락, 자음군 축소등과 같은 음운현상이 일어나면 결국 자음 하나만 남게 될 수도 있어서 그 자음이 마치 중첩과정에서 삽입된 것처럼 보일 수 있다.
Vaux (1998)의 분석에 의하면 강조형용사는 터키어에서 시작하여 몽골어, 퉁구스어와 같은 다른 알타이어로 확산해 나간 것으로 되어 있다. 비록 우리는 이에 대한 확신을 가지고 있지는 않으나 본 논문에서는 이를 지역확산(areal diffusion)의 예로 보고자 한다. 그 이유는 같은 현상이 알타이어군과는 친족관계에 있지 않지만 같은 지역에 속하는 아르메니아어에도 일어나고 있기 때문이다.
이상의 연구에서 얻은 결론은 최근에 유행하고 있는 최적성이론에 의한 중첩현상의 분석은 여러가지 면에서 알타이 어군에서 일어나는 강조형용사 현상의 적절한 설명에 부적절하다는 점이다.

The emphatic adjectives in Altaic languages are formed by reduplication of the first (C)V of the base followed by insertion of one of the consonants of {p, m, s, r}. In this paper, I reanalyze this peculiar reduplication in Turkish and Mongolian in an ...

The emphatic adjectives in Altaic languages are formed by reduplication of the first (C)V of the base followed by insertion of one of the consonants of {p, m, s, r}. In this paper, I reanalyze this peculiar reduplication in Turkish and Mongolian in an attempt to find its origin. The paper begins with analysis of Turkish adjectives, especially with the vowel initial bases which uniformly employ [p] as the linking consonant. Etymological examination of these examples reveals a possible origin of the linking consonant [p]. Etymologically speaking, [p]is not inserted in emphatic adjectives but it is the first consonant of the base, which is lost in word initial position in Altaic languages. For example, the first CV of Turkish uzun 'long' <*puzun is reduplicated as *pupuzun which then loses its initial [p] by the same rule, yielding upuzun. There is thus no insertion of [p] but simple CV reduplication. But synchronically it appears as if [p] had been inserted as part of the reduplication process. This synchronic [p] insertion is a result of grammaticalization of the original CV reduplication rule. The effect of this grammaticalization is that the same [p] sometimes appears even though the word did not begin with etymological [p], e.g. kara 'black' --> kapkara 'pitch-black'. It is possible that other linking consonants have a similar origin. Note, for example, in Turkish we have emphatic adjectives of the type ciplak 'naked' --> sirilciplak 'entirely naked', karisik 'confused' --> karmakarisik 'completely confused', which suggest possible origin of the linking consonants m and r. Interestingly these exampes remind us of Korean reduplicative constructions of the type khu- 'big' khu-na-khu-ta 'very big'; mel- 'far' me-na-mel-ta 'very far'; cha- 'cold' cha-ti-cha-ta 'very cold'; kil- 'long' kil-ko-kil-ta 'very long' The reduction of these constructions, by losing the medial vowel and simplification of consonant clusters, can eventually only leave a consonant, which on the surface may look as if it has been inserted as part of the emphatic reduplication process.
The above postulation of the origin of emphatic reduplication sheds some light on the following two questions: 1) Even though emphatic reduplication in Turkish is irregular in many respects, why is it that it is always [p] that is inserted when the base begins with a vowel and 2) Why, generally speaking, [p] appears to be elsewhere choice among the four linking consonants The answer is that [p] appears with vowel initial bases because originally vowel initial stems etymologically began with [p] and [p] is the elsewhere choice because of the grammaticalization process.
According to Vaux (1998), the emphatic reduplicaiton phenomenon has begun in Turkish and then spread to other Altaic languages such as Mongolian and Tungusic. Even though we do not know this for certain, it is a very likely scenario, given the fact that the phenomenon is most wide-spread in Turkic languages. In this paper, we analyze it as a case of areal diffusion of morphological process. The reason for this is that the same phenomenon also occurs in Armenian, which is an Indo-European language thus genetically unrelated to Altaic language family. But how the process actually spread remains to be investigated.
Finally, the paper compares this analysis with those done recently under the framework of Optimality Theory, especially the correspondence theory of reduplicaiton (McCarthy and Prince 1995). The conclusion that emerges from this comparison is that such synchronic analyses leave out many important questions unanswerd, such as 'why is it the case that [p] is always inserted when the base begins with a vowel ' The ramification is that we need to reexamine Saussure's synchronic/diachronic distinction, its implications, its consequences.

The emphatic adjectives in Altaic languages are formed by reduplication of the first (C)V of the base followed by insertion of one of the consonants of {p, m, s, r}. In this paper, I reanalyze this peculiar reduplication in Turkish and Mongolian in an ...

The emphatic adjectives in Altaic languages are formed by reduplication of the first (C)V of the base followed by insertion of one of the consonants of {p, m, s, r}. In this paper, I reanalyze this peculiar reduplication in Turkish and Mongolian in an attempt to find its origin. The paper begins with analysis of Turkish adjectives, especially with the vowel initial bases which uniformly employ [p] as the linking consonant. Etymological examination of these examples reveals a possible origin of the linking consonant [p]. Etymologically speaking, [p]is not inserted in emphatic adjectives but it is the first consonant of the base, which is lost in word initial position in Altaic languages. For example, the first CV of Turkish uzun 'long' <*puzun is reduplicated as *pupuzun which then loses its initial [p] by the same rule, yielding upuzun. There is thus no insertion of [p] but simple CV reduplication. But synchronically it appears as if [p] had been inserted as part of the reduplication process. This synchronic [p] insertion is a result of grammaticalization of the original CV reduplication rule. The effect of this grammaticalization is that the same [p] sometimes appears even though the word did not begin with etymological [p], e.g. kara 'black' --> kapkara 'pitch-black'. It is possible that other linking consonants have a similar origin. Note, for example, in Turkish we have emphatic adjectives of the type ciplak 'naked' --> sirilciplak 'entirely naked', karisik 'confused' --> karmakarisik 'completely confused', which suggest possible origin of the linking consonants m and r. Interestingly these exampes remind us of Korean reduplicative constructions of the type khu- 'big' khu-na-khu-ta 'very big'; mel- 'far' me-na-mel-ta 'very far'; cha- 'cold' cha-ti-cha-ta 'very cold'; kil- 'long' kil-ko-kil-ta 'very long' The reduction of these constructions, by losing the medial vowel and simplification of consonant clusters, can eventually only leave a consonant, which on the surface may look as if it has been inserted as part of the emphatic reduplication process.
The above postulation of the origin of emphatic reduplication sheds some light on the following two questions: 1) Even though emphatic reduplication in Turkish is irregular in many respects, why is it that it is always [p] that is inserted when the base begins with a vowel and 2) Why, generally speaking, [p] appears to be elsewhere choice among the four linking consonants The answer is that [p] appears with vowel initial bases because originally vowel initial stems etymologically began with [p] and [p] is the elsewhere choice because of the grammaticalization process.
According to Vaux (1998), the emphatic reduplicaiton phenomenon has begun in Turkish and then spread to other Altaic languages such as Mongolian and Tungusic. Even though we do not know this for certain, it is a very likely scenario, given the fact that the phenomenon is most wide-spread in Turkic languages. In this paper, we analyze it as a case of areal diffusion of morphological process. The reason for this is that the same phenomenon also occurs in Armenian, which is an Indo-European language thus genetically unrelated to Altaic language family. But how the process actually spread remains to be investigated.
Finally, the paper compares this analysis with those done recently under the framework of Optimality Theory, especially the correspondence theory of reduplicaiton (McCarthy and Prince 1995). The conclusion that emerges from this comparison is that such synchronic analyses leave out many important questions unanswerd, such as 'why is it the case that [p] is always inserted when the base begins with a vowel ' The ramification is that we need to reexamine Saussure's synchronic/diachronic distinction, its implications, its consequences.

연구보고서의 내용을 재정리하여 학술지에 발표후 그 결과를 현재 구상중인 저서 <Universal Phonological Processes: A coherent analysis of linguistic change>중 공시와 통시에 관한 부분에서 요약하여 소개할 예정임. 본 연구의 경험을 토대로 터키어, 몽고어, 한국 ...

연구보고서의 내용을 재정리하여 학술지에 발표후 그 결과를 현재 구상중인 저서 <Universal Phonological Processes: A coherent analysis of linguistic change>중 공시와 통시에 관한 부분에서 요약하여 소개할 예정임. 본 연구의 경험을 토대로 터키어, 몽고어, 한국어등 알타이 어족에 공통으로 나타나는 형태음운론적 현상의 비교분석에 더 매진할 계획임.