Abstract：This study examines word identity in Chinese disyllabic compounds through the perspective of character-word relationships, with a focus on coordinate compounds. Word identity refers to whether the same form (phonetic or written) represents one word or multiple words, or whether different forms represent the same lexical unit or distinct words. While previous studies have concentrated mainly on modern Chinese vocabulary, the identification of ancient compounds involves greater complexity due to evolving character-morpheme correspondences, phonological variation, and scribal conventions. This paper analyzes a series of historical compound groups to elucidate the mechanisms underlying the shifts between identity and non-identity.In disyllabic compounds, word identity is often reduced to morpheme identity. Imbalanced relationships between characters and morphemes—driven by variant graphs, phonetic loans, and cognate interchange—frequently generate multiple written forms for the same word. This research traces how graphic, phonetic, and semantic changes can alter lexical boundaries. The analysis first addresses the cases where variant forms represent the same word. For instance, “絜 (jie)” and “洁 (jie)” are functionally variant graphs representing the same morpheme. Although “絜” is the original form, “潔(洁)” emerged later and eventually predominated. “廉絜 (lianjie)” and “廉洁 (lianjie)” co-existed as variant forms of the identical lexical item. Another example is “録续 (luxu)” and “陆续 (luxu)”. The former represents the original phonetic-morphological form (録implying sequence/order). The character “陆 (lu)” functioned as a homophonic loan for “録 (lu)”, facilitated by the phonological convergence of the wu 屋, wo 沃, and zhu 烛 rhyme groups during the Tang and Song dynasties. Both forms represent the same coordinating compound meaning “in succession”.Lexical identity can emerge when originally distinct words, featuring a shared morpheme and a divergent morpheme represented by interchangeable graphs, functionally merge within specific semantic ranges. The compounds “专愚 (zhuanyu)” (stubborn ignorance) and “颛愚 (zhuanyu)” (naive simplicity) gradually merged in medieval texts, through character interchange between “专 (zhuan)” and “颛 (zhuan)”, yielding synonymous heterographic variants. Similarly, “瞻 (zhan)” (look towards) and “占 (zhan)” (divine/surmise; observe) are distinct yet likely etymologically linked morphemes. “瞻视 (zhanshi)” (looking/gazing) and “占视 (zhanshi)” (scrutinize for judgment/diagnosis) showed semantic overlap by the Six Dynasties period, as “占” was borrowed for “瞻” in certain contexts.Homophonic interference also plays a role in reshaping word identity. For example, “互 (hu)” and “护 (hu)” share the same pronunciation but represent distinct morphemes. “隐互 (yinhu)” and “隐护 (yinhu)” differ significantly in lexical and grammatical meaning. The former describes mathematical complexity (“interlaced concealment”), textual obscurity, or intricate patterns, while the latter means“conceal physically” or “shield (wrongs/thieves)”. While “隐互” primarily functions as an adjective and occasionally as a causative verb, “隐护” exclusively serves as a verb. Ming Dynasty literature contains examples where “隐互” is used in place of “隐护” to mean “shelter” or “concealment”. Such substitutions may be attributed to occasional phonological borrowing.Conversely, variant forms of the same word can produce misinterpretations unrelated to its original meaning and eventually be treated as different words. For example, “檃栝 (yinkuo)” is originally a single word meaning a carpenter’s frame for straightening wood, metaphorically “correct/conform” or “adapt (text)”. Graphic confusion between “栝” and “括” was common. In medieval times, “隐括 (yinkuo)” developed a new, disconnected meaning of “scrutinize/investigate”. Later, another unrelated meaning of “summarize/generalize” emerged. These novel meanings, arising from graphic-form interference and lacking etymological connection to the original sense, must be considered distinct lexemes from earlier “檃栝”. Likewise, the transition from “擗踊 (piyong)” to “躃踊 (piyong)” initially involved only a radical change, yielding a variant form of the same word. However, the graphic change obscured the compound’s bimorphemic nature (“chest-beating” and “foot-stamping”). Subsequently, “躃踊” semantically narrowed, losing the chest-beating component. “躃” developed meanings like “stumble/fall” and “躃踊” came to signify actions like leaping/jumping and falling down. The two forms have thereby become homophones with different morphemes.In conclusion, the interface between character and morpheme is decisive for determining the identity of disyllabic words. Asymmetries in this mapping require contextual analysis to distinguish variant representations of the same word from genuinely distinct lexemes. Diachronic fluidity permits bidirectional shifts between identity and non-identity: phonetic loans and graphic variation often generate alternative forms, while scribal errors and folk etymology can split or reconfigure semantic boundaries. These findings provide practical implications for Chinese historical lexicology. Future corpus-based research can further refine the identification and interpretation of ancient compounds.
王诚. 试论字词关系视角下汉语复合词的“同一性”问题[J]. 浙江大学学报(人文社会科学版), 2026, 56(8): 29-42.
Wang Cheng. Word Identity in Chinese Compounds from a Character-Word Relationship Perspective: A Case Study of Coordinate Disyllabic Words. JOURNAL OF ZHEJIANG UNIVERSITY, 2026, 56(8): 29-42.
2026, Vol. 56 
