“Sound symbolism is the direct linkage between sound and meaning” (Hinton et al. 1994), a phenomenon usually introduced as exception to the Saussurean thesis of the “arbitrariness” of linguistic signs. The direct origin of modern research is usually considered to be the 1920s. A lot of early experimental studies on sound symbolism stated the association of sounds with shapes and arrived at a positive conclusion on the existence of (universal) sound symbolism. Frequently cited examples include onomatopoeic forms and antonymic adjectives. The meanings of these words are more or less imaginable from their sounds when properly highlighted (e.g., presented in pairs or in context). Sound symbolism has attracted many key scholars in modern linguistics and psychology: whereas psychological experiments occupy a large part of the literature, linguists have also conducted descriptive investigations in normal lexical items by means of vocabulary lists. This second approach has gained statistical reinforcement from recent quantitative studies. Thus, the large majority of studies agree that languages involve sound symbolism, and speakers of the languages can more or less sense this. When outlining both traditional and new studies on sound symbolism, and their main descriptive findings, special attention is paid to ideophones, “imitative” words including, but not restricted to, onomatopoeia. This paper then indicates a core issue regarding the ambiguity in the interpretations of both “meaning” and “sound” in sound symbolism. The two types of ambiguity are attributed to the hierarchical nature of meaning and sound. This paper further considers possible motivations and mechanisms for sound symbolism, focused on in recent debates, addressing three major issues as to how and why sound symbolism occurs: its articulatory and acoustic bases, its universality and language-specificity, and the nature-nurture question. Sound symbolism appears an intricate phenomenon, a comprehensive grasp of which will require an interdisciplinary approach, which has been critically lacking in literature. Especially technological advancements in (neuro-)psychological studies will benefit from the growing linguistic study of ideophone typology. Sound symbolism research has a century-long history, and various sound-meaning mappings have been described in (speakers of) various languages. These findings need reinterpretation in terms of the levels of semantic and phonological generalization, and the reinterpretation will offer a sound basis for an extensive comprehension of the motivations and mechanisms of sound symbolism under debate. Two trends of practical studies indicate the educational applicability of sound symbolism and ideophones, and confirm the crucial role of sound symbolism in naming.