The term phonestheme (//; phonaestheme in British English) was coined in 1930 by British linguist J. R. Firth (from the Greek φωνή phone, "sound", and αἴσθημα aisthema, "perception" from αίσθάνομαι aisthanomai, "I perceive") to label the systematic pairing of form and meaning in a language. Such pairing would violate the arbitrariness principle of semantics.
Phonesthemes are of critical interest to students of the internal structure of words because they appear to be a case where the internal structure of the word is non-compositional; i.e., a word with a phonestheme in it has other material in it that is not itself a morpheme. Phonesthemes "fascinate some linguists", as Ben Zimmer has phrased it, in a process that can become "mystical" or "unscientific".
For example, the English phonestheme "gl-" occurs in a large number of words relating to light or vision, like "glitter", "glisten", "glow", "gleam", "glare", "glint", "glimmer", "gloss", and so on; yet, despite this, the remainder of each word is not itself a phonestheme (i.e., a pairing of form and meaning); i.e., "-isten", "-ow", and "-eam" do not make meaningful contributions to "glisten", "glow", and "gleam". There are three main ways in which phonesthemes are empirically identified.
The first is through corpus studies, where the words of a language are subjected to statistical analysis, and the particular form-meaning pairing, or phonestheme, is shown to constitute a statistically unexpected distribution in the lexicon or not.
Corpus studies can inform a researcher about the current state of the lexicon, a critical first step, but importantly are completely uninformative when it comes to questions of whether and how phonesthemes are represented in the minds of language users.
Study of patterns in neologisms
The second type of approach makes use of the tendency for phonesthemes to participate in the coinage and interpretation of neologisms (i.e., new words in a language). Various studies have demonstrated that, when asked to invent or interpret new words, subjects tend to follow the patterns that are predicted by the phonesthemes in their language. It is known, for example, that the word bangle is a loan from Hindi but speakers tend to associate it with English onomatopoeia like bang. While this approach demonstrates the vitality of phonesthemic patterns, it does not provide any evidence about whether (or how) phonesthemes are represented in the minds of speaker-hearers.
Study of linguistic processing patterns
The final type of evidence uses the methods of psycholinguistics to study exactly how phonesthemes participate in language processing. One such method is phonesthemic priming — akin to morphological priming — which demonstrates that people represent phonesthemes much as they do typical morphemes, despite the fact that phonesthemes are non-compositional.
Discussions of phonesthesia are often grouped with other phenomena under the rubric of sound symbolism.
While phonesthemes have mostly been identified in the onsets of words and syllables, they can have other forms. There has been some argument that sequences like "-ash" and "-ack" in English also serve as phonesthemes, due to their patterning in words that denote forceful, destructive contact ("smash", "crash", "bash", etc.) and abrupt contact ("smack", "whack", "crack", etc.), respectively.
In addition to the distribution of phonesthemes, linguists consider their motivation. In some cases, there may appear to be good sound-symbolic reasons why phonesthemes would have the form they have. In the case of "-ack", for example, we might imagine that the words sharing this phonestheme do so because they denote events that would produce a similar sound. But critically, there are many phonesthemes for which there can be no sound-symbolic basis, such as "gl-", for the simple reason that their meanings (such as 'pertaining to light or vision') entail no sound.
While there are numerous studies on living languages, research is lacking about ancient languages, although the first documented example of phonesthemes dates back to at least the fourth century B.C.: Plato's Cratylus clearly mentioned a gl- phonestheme (a different one from that discussed previously, as those words are not of Greek origin) as well as an st- one and gave an explanation in terms of phonosemantics.
Examples of phonesthemes in English (aside from "gl-"), include "sn-", related to the mouth or nose, as in "snarl", "snout", "snicker", "snack", and so on; and there is also a set of words starting with "sl-", which appears in words denoting frictionless motion, like "slide", "slick", "sled", and so on - which are themselves a subset of a larger set of words beginning with sl- that are pejorative behaviours, traits, or events: slack, slouch, sludge, slime, slosh, slash, sloppy, slug, sluggard, slattern, slut, slang, sly, slither, slow, sloth, sleepy, sleet, slip, slipshod, slope, slit, slay, sleek, slant, slovenly, slab, slap, slough, slum, slump, slobber, slaver, slur, slog, slate.
- "Phonestheme". Merriam-Webster Dictionary.
- Firth, J. R. (1964) . The Tongues of Men, and Speech. London: Oxford University Press. p. 211.
- Bob Garfield and Mike Vuolo (2014-10-20). "Where Did the Word Snark Come From?". Lexicon Valley (Podcast). Slate. Event occurs at 0:27:30. Retrieved 2015-12-06.
- Bolinger, Dwight (1980). Language, the Loaded Weapon: The use and abuse of language today. New York: Longman. p. 20. ISBN 0582291089.
- Leonardi, Filippo Maria (2015). Phonesthemes in Latin Language. Academia.edu.
- Marchand, Hans (1959), "Phonetic Symbolism in English Word-Formation", Indogermanische Forschungen, Volume 64, pp.146-168, 256-277 , 
- Marchand, Hans, The Categories and Types of Present-Day English Word Formation: A Synchronic-Diachronic Approach, 2nd ed, C.H.Beck'she Verlagsbuchhandlung, (München), 1969. Chapter 7 ("Phonetic symbolism") and chapter 8 ("Motivation by linguistic form: Ablaut and rime combinations"), pp. 397–439.
- Leonardi, Filippo Maria (2015), "Phonesthemes in Latin language"