혼성어는 두 단어에서 각각 일부를 잘라내고 결합하여 만들어진 단어이다. 절단과 합성이라는 두 가지 기제가 관여한다는 점에서 절단어나 합성어와 구별된다. 혼성어(예: 레캉스, 라볶이)는 합성 이전의 두 단어가 갖는 지시 대상의 속성을 모두 갖는다는 점에서 두음절어(예: 노찾사, 개콘)와도 구별된다. 본고는 혼성어를 계열적 혼성어와 통합적 혼성어로 구별하여 살펴보고 혼성어의 음절수 제약에 중점을 두고 혼성어 형성 과정을 고찰하였다. 계열적 혼성어는 혼성되는 두 단어가 동일한 위치에 출현할 수 있어 분포가 같으며 유사한 의미장에 속하는 단어들 사이에서 일어난다(예: 레캉스, 개그운서). 통합적 혼성어는 혼성되는 두 단어가 같이 출현하여 구를 형성할 수 있으며 두 단어는 주로 수식 관계가 될 수 있다(예: 유티즌, 모티켓). 혼성어 형성시 두 단어가 절단되는 부분에 따라 (1) AB＋CD→AD형: 개그맨＋아나운서→개그운서, (2) AB＋CD→ABD형: 김치＋할리우드→김치우드, (3) AB＋CD→ACD형: 컴퓨터＋도사→컴도사, (4) AB＋CD→BD형: 아줌마＋신데렐라→줌마렐라 등으로 나눠 살펴보았다. (1) AD형과 (4) BD형은 두 단어에서 절단이 일어나 전체 혼성어의 의미가 불투명하다. 또한 후행 단어의 음절수에 기반해서 혼성어가 형성된다는 점에서 전형적인 혼성어로 볼 수 있다. (2) ABD형과 (3) ACD형은 한 단어에서만 절단이 일어난 경우이다. (2) ABD형은 혼성어가 처음 형성될 때는 ‘김치우드’처럼 음절수 제약을 지키게 된다. 그러나 후행요소 D가 ‘텔, 파라치’처럼 접사화되면 음절수 제약도 없어지고 파생과 유사해진다. (3) ACD형은 ‘컴도사’와 같이 후행 단어가 절단되지 않으므로 음절수 제약이 없다. 동시에 후행 단어에 기반해서 혼성어가 형성되지 않으며 의미도 어느 정도 투명하여 전형적인 혼성어로 보기 힘들다. 더구나 ‘컴’ 등이 단어의 지위를 얻으면서 혼성어 고유의 특성이 상실되는 경향이 있다.

Blends are words formed by combination of arbitrary parts of existing two words. Blends are characterized by comprising both clipping and compounding. Blends share the references of both existing words(ex. 레캉스, 라볶이). In this respect, they are differentiated from acronyms(ex. 노찾사, 개콘). This paper classified blends into paradigmatic ones and syntagmatic ones, and examined the blending process in terms of the syllable constraint. In paradigmatic blends, two existing words have the same distribution and occur between the words that could be included in the same semantic field(ex. 레캉스, 개그운서). In syntagmatic blends, two existing words can form a phrase and stand in the modifying relation(ex. 유티즌, 모티켓). According to the way how two existing words are clipped, blends are classified into four patterns. (1) AB＋CD→AD pattern: 개그맨＋아나운서→개그운서, (2) AB＋CD→ABD pattern: 김치＋할리우드→김치우드, (3) AB＋CD→ACD pattern: 컴퓨터＋도사→컴도사, (4) AB＋CD→BD pattern: 아줌마＋신데렐라→줌마렐라. In the patterns AD and BD, clippings occur in both words, and the meaning of blends are opaque. These blends are typical in that they are formed on the basis of the number of syllables. In ABD and ACD patterns, clipping occur only in the one word. The ABD pattern abides by the syllable constraint when forming the blends like ‘김치우드’. But when the following elements begin to obtain the affix status, this constraint disappears and this pattern resembles the process of derivation. The ACD pattern doesn’t have the syllable constraint because the following element doesn’t undergo the clipping process like ‘컴도사(컴퓨터+도사)’. At the same time, the blends aren't based on the following word, and the meaning of the blends are predictable. In these respects, this pattern isn’t a typical blend. Besides according as the clipped parts such as ‘컴’ obtain the word status, the property of blending tend to fade away.

Blends are words formed by combination of arbitrary parts of existing two words. Blends are characterized by comprising both clipping and compounding. Blends share the references of both existing words(ex. 레캉스, 라볶이). In this respect, they are differentiated from acronyms(ex. 노찾사, 개콘). This paper classified blends into paradigmatic ones and syntagmatic ones, and examined the blending process in terms of the syllable constraint. In paradigmatic blends, two existing words have the same distribution and occur between the words that could be included in the same semantic field(ex. 레캉스, 개그운서). In syntagmatic blends, two existing words can form a phrase and stand in the modifying relation(ex. 유티즌, 모티켓). According to the way how two existing words are clipped, blends are classified into four patterns. (1) AB＋CD→AD pattern: 개그맨＋아나운서→개그운서, (2) AB＋CD→ABD pattern: 김치＋할리우드→김치우드, (3) AB＋CD→ACD pattern: 컴퓨터＋도사→컴도사, (4) AB＋CD→BD pattern: 아줌마＋신데렐라→줌마렐라. In the patterns AD and BD, clippings occur in both words, and the meaning of blends are opaque. These blends are typical in that they are formed on the basis of the number of syllables. In ABD and ACD patterns, clipping occur only in the one word. The ABD pattern abides by the syllable constraint when forming the blends like ‘김치우드’. But when the following elements begin to obtain the affix status, this constraint disappears and this pattern resembles the process of derivation. The ACD pattern doesn’t have the syllable constraint because the following element doesn’t undergo the clipping process like ‘컴도사(컴퓨터+도사)’. At the same time, the blends aren't based on the following word, and the meaning of the blends are predictable. In these respects, this pattern isn’t a typical blend. Besides according as the clipped parts such as ‘컴’ obtain the word status, the property of blending tend to fade away.

, , , , , ,

blends, word-formation, clipping, compounding, acronyms, the number of syllable, semantic transparency