Skip to navigation – Site map

The submorphemic conjecture in English: towards a distributed model of the cognitive dynamics of submorphemes

Didier Bottineau


This study is based on a general review of the vowel and consonant clusters known as submorphemes, ideophones and phonaesthemes that are evidenced in the lexicon of the English language. The aim is to investigate the cognitive relevance of submorphology for language users and observers and try to determine whether such relevance exists at all or whether the linguist is the only one to perceive a link between submorphemic order and lexical-semantic classes. Various aspects of the question will be explored: the nature of submorphemic semantics (specifically in the case of English) – expressive, impressive or mechanistic (classifying lexical items and catalyzing form / matter connections in both directions) and their differentiated distribution (the speaker’s, hearer’s and linguist’s respective experiences of submorphemics).

Top of page

Full text


1The question of lexical submorphemics has long been controversial in English lexical studies and in general (Genette 1976, Jakobson 1980, Toussaint 1983). The linguistic sign is traditionally considered as “arbitrary”: the phonological form is not iconic and does not reflect the physical properties of the object or event to which the word refers; for lack of motivation there are as many words as languages for one given item (snow, neige in French, elur in Basque, erc’h in Breton) and no human community produces “ideal” signs as is accomplished by Adam in the book of Genesis or by the logothete according to the Cratylus. On the other hand, a minority of authors are convinced that some areas of the English lexicon are organized by a paradigm of submorphemic markers: clusters of more than one phoneme, and occasionally isolated phonemes, which appear to relate the lexical notion to a specific semantic field.

2Between these two positions there is hardly any room for scientific debate. Submorphologists usually accumulate impressive bodies of data (Nemo for French, Bohas for Arabic and other Semitic languages) which are interpreted in the light of elaborate theoretical models such as Bohas’s Matrices, Etymons and Racines and Philps’s sublexical markers which are described in their synchronic, diachronic and panchronic dimensions; but the sceptical opponents consider that the undertaking is illegitimate and ill-targeted in the first place so that the accumulated data and analyses, far from supporting the cause of submorphologists, is at best ignored and occasionally taken to illustrate the extent to which one can err. The gap separating the two sides could hardly be wider; for a productive scientific debate to be inaugurated, what is urgently needed is to reconcile antagonistic views over at least one issue: the real nature of the problem.

1. Establishing the issues: towards an embodied and distributed conception of speech and the word

3At the root of the misunderstanding lies the fact that the questions for which those positions are supposed to provide an answer have not really been raised in the first place. What is, after all, a word? Basically, a segment of vocal action undertaken by a human subject, usually in the presence of other human subjects. In a series of recent studies the author of these lines has proposed a model of the speaking activity that can be summarized as follows (Bottineau 2008).

4Vocal speech takes place when a human subject starts altering the acoustic properties of the atmospheric environment. This perturbation is perceivable by all embodied consciences present, including the emitter’s own.

5Utterances, whether they take the form of fully-fledged standardized sentences or of partial and improvised sequences, will have a distributed cognitive impact on all brains involved: interpretation. To interpret is (i) to realize a semantic construction on the basis of the sequence of vocal keys received and (ii) to map the semantic responses against the pre-existing psychological context characterizing each individual subject (personal generic background and specific knowledge, moods and emotions, intentions, social relations and so on). Because interpretive dynamics is based on the perceptions of the same signal the outcomes will tend to converge, and because the psychological contexts enacting them are different interpretive dynamics is inevitably divergent. Verbal constructivism is about the emergence of concerted semantic and pragmatic projects out of psychological diversity and idiosyncracies.

  • 1 This metaphorical, somewhat Californian-sounding terminology is only meant to emphasize the collabo (...)

6Speech should not be reduced to communication: speech is also used in the process of forming one’s own conscious ideas like Why is that silly dog barking? Or I must be dreaming. The same vocal ritual is used for both transitive vocal speech and reflexive inhibited speech: a speaker will orient vocally (in Maturana’s sense) a hearer’s interpretive mind map in the same way as a thinker will orient silently his own sense-making itinerary; to think is to anticipate what the auditory and tactile proprioception of one’s own voicing would sound and feel like if it were actually carried out motorically, and internal discourse does indeed provide evidence of some of the prosodic features of actually vocalized speech (rhythm, word and sentence stress, tone units, melody, etc.). The same orienting process is used to direct either an external conscience’s activity (communication, with intended meaning), or the private activity of one’s own internal conscience (meditation, generating unintended semantic emergence); for this reason we will borrow from Maturana the term orientee to designate both the external interpreter (an actual hearer) and the internal one (the reflexive consciousness experiencing the semantic benefit of its own simulated vocal process). We are therefore rejecting ‘encodingism’: even if language is commonly used to express ideas (or so we believe – and how could we ever talk without this necessary faith?), the effect of the chain of vocal action is not to organize and give acoustic substance to the externalization of internally formed representations, but to conduct by vocal motricity the distributed procedure of semantic networking1. In this perspective, the linguist’s empirical objects of investigation (the lexicon, morphology, syntax and prosody) are to be redefined in terms of vocal sub-actions, each of them yielding a specific cognitive contribution to the general orchestration of the sense-making process.

7Syntax: Word order cannot be explained in terms of information packaging (Chafe 1976). Syntax, or the distinction and ordering of the successive steps of the sense-making process (both transitive and reflexive), is about the prototypical mind-map characterizing ideation in a given language; to this day, three different cognitive types have been proposed, but they cannot be recalled here.

8The lexicon: words are ritual vocal actions whose perceptual-motoric coupling is imposed by the surrounding speaking community of local experts (parents, relatives, teachers, friends) on developing and language-forming children in the presence of heterogeneous but consistent and regular sets of situations and in the context of stereotypical discourse (Putnam 1975, Anscombre 2001). For one given individual subject, the meaning of the word is a network of synchronized and correlated heterogeneous memories of impressions, both non-linguistic (Culioli’s notion with all its pragmatic and cultural dimensions) and verbal (the sequences of discourse in which a word is regularly encountered, both formally and thematically: Anscombre’s stereotypes) – dog is a word I and others use in a whole range of contexts (including in the absence of the dog) and each reiterated occurrence of the word will map the recorded hierarchized network of previously experienced impressions against the other structurally similar networks reactivated by other lexical units, to form a consistent hyper-network: a mental scene. Abstraction is not a problem: there does not have to be an empirical common denominator for a given word (take abstraction, for example) apart from the type of social situations and discourse contexts in which it is ordinarily used. “A word is a powerful madeleine” (Stéphane Robert 2003): a constant and reiterated piece of vocal action linking heterogeneous sections of experience with converging and diverging properties; the role of this socially controlled motoric token is to map, when needed, a given experiential occurrence against the knowledge and pragmatic know-how acquired in the course of all previous occurrences of the word / real life coupling: the word maps singularize currently experienced occurrences of categories against a recorded background of features retrieved from multiple previous occurrences; literally, language amplifies intelligence (in the etymological sense): vocal action catalyzes the retrieval of previously stored knowledge, a retrieval which could not be executed so efficiently and extensively without the socially-nurtured tokens. This vocal intelligence agency is used as a guideline to form predictions and make decisions, that is, promote action. As the lexicon is imposed on the child by its speaking kin, as speech shapes dynamically the acoustic environment in which the infant grows and to which its development adapts, so words act as the social and cultural profilers of the category-making processes: thanks to these social agents of vocal intelligence, personal memories and experience are framed by externally imposed labels and contrastive networks (dogs as opposed to cats in a given cultural-linguistic collective way of life).

9Morphology: just as notional words are vocal sub-actions that activate recorded networks of impressions, functional morphemes are vocal sub-actions of a different kind: they activate recorded patterns of connections whose effect is to relate notions to one another and to the currently experienced situation in terms of time, space, determination and so on; the outcome is the formation of the hyper-network required to achieve a “mental scene” that is correlated with the ongoing experience, both material and psychological, both verbal and non-verbal.

10Prosody: melody and rhythm in English are mainly used to plan and fine-tune the reception of the ongoing orienting process in the intersubjective interplay; but it will be affected in real time by the emotional effect of the visual perception of the addressee’s and witnesses’ reactions.

11To summarize, speech is to be understood as an individual vocal process whose effect is to coordinate the cognitive activities of all receptors involved by focussing them all on the same sense-making procedure, to be mapped against contrasted psychological contexts and backgrounds. The coordinative span is restricted to the individual self when the vocal emission of the acoustic signal is inhibited, and extended to all the attainable subjects otherwise (including the use of technological artefacts to lengthen the perceptual reach of the signal and record or stabilize it: Bruner’s amplifiers). The vocal speaker is the individual director of the cognitive coordination, and dialogue consists in ‘voicing over’ the subjective vocal leadership of the coordinative process in a constructivistic way. In the debate opposing internalism and externalism, this account is a moderate compromise: the biological loci of the verbally generated mental scenes are the nervous systems confined in the cortices, but the neural dynamics through which the semantic constructions are carried out involves the motoric activity of the whole body’s muscles and its perception by the nervous system. Linguistic cognition is not confined within neural systems, but distributed; the whole metabolism is part and parcel of cognitive dynamics, which are embodied, and not in metaphorical and representational terms (which implies a rejection of some of the major tenets of the cognitivist paradigm). And as this motoric activity affects the atmospheric mediating environment in a perceivable way for all bodies present, linguistic cognition cannot be individual except when its muscular motricity is inhibited (internal meditative thinking): at this second level, cognitive processes are distributed over bodies, brains and the environment.

12The coordinative dynamics involves a plurality of agents in a given medium and will form a fluctuating collective and concerted mind. The anthropological effects are well known: the human species is the only one to have developed an aptitude for subjectively-led collective self-determination (to be opposed to bees’ and ants’ functionally rigid systems of pragmatic coordination). As a result, it is also the only species to have self-determined and concerted its relation to the other species, taking over the whole environment and turning it into its own private ecological niche – a process that started with sedentarization and agriculture and is now reaching a climax with rampant demography, the globalization of exchanges in all their forms, the fall of biodiversity, and climate change.

13In a nutshell, languaging is about the reciprocal effects of neural and muscular ritual motoric actions learned by the developing subject in the social and cultural environment. This conception paves the way for a more technical formulation of the submorphemic conjecture. Paradoxically, there is no difference in nature between the signifier and the signified: the lexical signifier is a vocal sub-action, a multimodal motoric-perceptual coupling and programme to be inserted in a hyperordinate sequence of analogous sub-actions; and the lexical signified is a network of recorded impressions experienced in the course of real life, that is, another network of multimodal motoric-perceptual couplings, both verbal and non-verbal. The difference is that the signifier is a relatively stable and consensual programme of vocal action, while the galaxy of experiences out of which the collected impressions forming the features to be organized into notional networks emerge is largely chaotic, despite the streamlining effect of parenting, schooling and socializing in general. A theory of the lexicon is a model of the linking of those two types of perceptual-motoric couplings: the vocal token, the signifier; and the varying collection of impressions attached to it, the signified. A theory of phonetic change is a model of how inevitable idiosyncratic ‘errors of phonatory copy’ in the analogical reproduction of sounds are made to converge and form strains of phonetic evolutions to which specific sub-groups may identify (or emerge) and compete with other communities. And a theory of semantic change is a model of how the random differences in the idiosyncratic collections of impressive items forming notional networks are made to converge and form emerging strains of interpretation.

14The submorphemic conjecture consists in hypothesizing that in English the word is not always the smallest relevant vocal sub-action to be singled out in the monitoring of the distributed sense-making process; that submorphemic markers constitute, as it were, hypo-sub-actions whose specific orienting effect can be modelled. To explore this hypothesis, the following sections are devoted to

  • a review of some of the arguments against the establishment of submorphemic markers as a relevant object worthy of scientific scrutiny;

  • the presentation of the nature of the semantics of submorphemic markers: their orienting effect;

  • an exploration of the ways in which vocal “hypo-sub-acting” may be acquired by developing infants and of the traces their popular intuitive awareness leaves in some adult language practices.

2. Establishing the object

2.1. Submorphemes are best described by the following objective properties

(i) a [submorpheme] is an identifiable segment in a {lexical unit}: {[st]and}, {sh[amble]}, {[st][ump]}.

(ii) to the consenting linguist, this segment appears as a common submorphological denominator shared by a list of lexical units belonging to the same semantic field: in the semantic field of ‘stativity’ (in space or time) and rigidity (of an object) we find still stall stand stay stare, along with many others. The following table presents some of the most often studied initial consonant clusters:

Fig. 1: Some submorphemic consonant clusters in the lexicon of English

Consonant cluster

Semantic class

Location within the syllable
(onset, coda, both)

+ R


Centrifugal rotation

(operation), and/or

projection (result)

spin span spill speak spew spit spend speck spot spate spall spawn spook spool spam spoon

clasp wisp

sip seep sap soap soup

spray sprawl sprinkle spring sprightly sport


Stativity or fixity

in space or time

still stall stand stay stare start stop step stab stump stumble star

rest mast cast thrust fist bust

sit set site

stray string strewn stretch



(two-dimensional object,

and/or movement

applied to it

skin skull skate skid skittle scamper scab

mask cask


Scrooge scour screech scratch scrawl scribble scrofula scrub






wrought wrath writhe wrist


oscillation, pendulation

swing swoon sway swoop swear switch sweep swap swagger


sow sew




two twin -tween twitch




CL (gl)

clinging, promiscuity

cling clench clasp clutter glue clutch clog



wriggle wrangle haggle

GL (cl)


glint glisten glitter glow glory glimmer glee





non-vertical movements

sleep slay slope sleet slot slate sling slug



jostle wrestle

(iii) concerning their structure, submorphemes are formatted by the morphophonological constraint. According to Tournier (1985) they may cover either the onset of the syllable (stand), or the rhyme, which must possess a coda (shamble).

15In the onset, submorphemes are usually bi- or tri-consonantal clusters ([st]and, [str]etch) but analyses will vary. Among the 29 initial submorphemes he postulates, Tournier admits a limited number of uniconsonantal submorphemes: w for instability, oscillation as in wag, waddle, wobble; v for violence as in violent, vicious, vituperate, virulent. As regards bi- and tri-consonantal clusters, Bottineau (2002, 2003a) has suggested that three tri-consonantal submorphemes, spr, str and scr, are actually variants of sp, st and sk complemented by an additional r related to the notion of dynamic movement and agentivity. It should be added that apart from scl- (sclerosis, skeleton) no other format for tri-consonantal clusters is available in the range of possibilities offered by the morphophonological constraint in English (*stl, *swr etc.): it is usually considered that English, unlike Semitic languages, compresses its submorphemes in the form of clusters located in one of the pre- or post-vocal locations in the syllable: the onset, the coda. However the question whether a distributed approach to submorphology should not be preferred sometimes emerges: should discontinuous matrices be admitted? This might be the case for sit, sat and set in the case of s-t, and other instances of discontinuous matrices have long been identified: loop, lap, lip, and maybe leap; full and fill. If it were so, tri-consonantal matrices are potential candidates for submorphology: spin, span, spoon, spank, spend. It is worth noticing that this question echoes a current debate among specialists of the Semitic languages: for traditional scholars, matrices are tri-consonantal, while for contemporary scholars, they are bi-consonantal matrices to which one consonant is added (special mention must be made here of Bohas’s more elaborate model). Moreover, Zev bar-Lev (2008) postulates the existence of key consonants in Arabic in initial position within a marginal uni-consonantal model. The lexicon of English, however, is primarily based on one-syllable units, leaving little space for tri-consonantal matrices in the system.

16For the rhyme and the coda, Tournier itemizes nine submorphemes (including some cases of vowel alternation), once again uni-, bi- and tri-consonantal (VC1>3). Some submorphemic markers may appear either in the onset or in the coda if the morphophonological constraint will allow it in terms of the number and location of mores: spend, wisp; skin, mask; stick, mast. In some cases both the onset and the coda host submorphemic consonant clusters: flask, clasp. Initial submorphemes are semantically more salient than final ones: the former select a primary classifying property while the latter provide a secondary, complementary feature. To substantiate this with arguments other than subjective semantic intuition (or aspiration) a dynamic theory of the syllable is required. In Diver’s Columbia School of Phonology (Phonology as Human Behaviour, Tobin 1997) the vowel and the consonant are not sets of static features but dynamic actions: vowels are obtained by the process of lowering the lower jaw (increasing the aperture) from a minimum (closed /i/) to a maximum (open /a/), while consonants are obtained by the process of raising the same organ (reducing the aperture) from a minimum (approximate) to a maximum (stops, with total stricture) through fricatives (obstruents). Combining this proposal with the CVC syllabic format of the morphophonological constraint entails that the syllable is a unit of expiratory action during which up to three mandibular shifts may occur (closing, opening and closing), forming the syllable’s kinetic cyclical programme (cf. MacNeilage’s Frame/Content Theory, 1998). Recent studies suggest that the syllable has emerged from recombining kinetic actions that pre-existed separately: shouting (an alteration of expiring by the vocal chords) and, for the mandibular oscillation, eating in all its forms (breastfeeding, licking and so on). In this view, the CVC template is the integration and coordination of a mandibular oscillation borrowed from a specific family of motoric programmes with an expiratory sequence, belonging to another family of programmes. The shift from a symbolic approach to an embodied one requires that these elements be taken into account.

2.2. Arguments against submorphemics

17Submorphemes are usually considered as non-existent objects for a number of reasons, including the following:

  • submorphemes are unnecessary and therefore sporadic: birds of a feather flock together – in this proverb the verb flock can be related with one submarker, fl-, but the noun bird does not appear to contain any such element, and yet manages to convey an ‘avian’ notion quite successfully, despite the absence of any classifying submorpheme.

  • submorphemes are not unified and continuous linguistic objects: what may appear as objects in present-day English cannot always be traced back as continuous objects in the history of the language. Philps (2003b) has extensively discussed the role of Indo-European *s- mobile in the formation of bi-consonantal initial clusters. The inclusion of one given lexical item depends on the semantic definition of the submorpheme in the first place: if sp is defined as “pointed”, as is done by Tournier, spend is to be excluded from the series. However, Bottineau proposes “a rapid centrifugal movement”, which includes pointed objects that can be cast by means of a similar movement (spear, spike). In this definition spend is included (cf. to go on a spending spree), but this in turn raises the question of the continuity of the submorpheme sp in lexical diachrony: the present-day paradigm, which includes spear and spend, is made to incorporate words bearing an original sp- cluster along with words historically starting with a Latin ex- prefix; for this reason, submorphology is frequently rejected as folk etymology.

18Discussion: this argument mixes two dimensions, synchrony and diachrony. Synchrony is about how a subject is able to construct a systematic know-how out of reiterated experiences, both passive (interpretation of others’ statements) and active (personal production). It is difficult to see how this know-how may take into account the academic knowledge of the word’s history, as no popular access to it is can be observed. Instead, analogy may be instrumental in fostering morpho-semantic connections: random resemblances between semantically similar words are taken as memo-technical cues and become submorphologically meaningful; thus a word which evolves phonologically by reducing a prefix to an s- incidentally becomes a competent candidate for incorporation into the ‘club’ of sp-class nouns, and, if the connection turns out to appear relevant, even metaphorically, the inclusion is implemented.

  • submorphemes have too many exceptions. This in fact depends on two factors: firstly, the semantic categorization of the lexical class; see above the discrepancy between Tournier and Bottineau concerning sp: the latter’s formulation makes sp’s inclusive power much more far-reaching, and secondly, the formal properties of the submorpheme considered, as well as, more generally the boundaries of the morphological format accepted for submorphemes. As a rule, initial bi- and tri-consonantal clusters are those with the lowest proportion of exceptions, as low as 10% for sp, although quantifying is problematic on account of derivation and frequency. If one considers Tournier’s uni-consonantal clusters, more exceptions than matches are likely to be found. As regards final submorphemes, they are more limited in number and should probably be treated as a particular case. Furthermore, scores of ordinary words like car, chair, eat are devoid of submorphological semantic classifiers: obviously, the system is not intrinsically needed; not all words bear the morphological signature of the semantic field they belong to. In other words, up to a certain point, the submorphology of the English lexicon bears some resemblance to the semantic classifiers of Bantu and West-Atlantic languages, but without their formal systematicity.

    • 2 The teeth are used as passive articulators against the tongue, but as active articulators against t (...)
    • 3 In Stanley Kubrick’s film 2001: A Space Odyssey, the computer Karl manages to ‘interpret’ the astro (...)

    submorphemes are phonosymbolic. Clarifications are required here. (i) literally, any morpheme is phonosymbolic by definition: a piece of vocal action like chair is supposed to be reminiscent of the prototypical object in the presence of which the word was repeatedly heard by the infant in the presence of adults. Literally, any meaningful piece of vocal action is an ideophone. (ii) submorphemes are sensomimetic: on the semantic side, consonant clusters underline some piece of sensori-motor experience through which an object is apprehended: st points to a notable lack of movement, sp to a centrifugal movement (according to Bottineau), scr to a friction, gl to luminosity, and so on. The difficulty is that on the phonic side, consonant clusters are also pieces of sensori-motor experience: vocal hypo-sub-actions involving the interaction of active articulators (the vocal cords, the blade, the apex, the teeth, the lips) and passive articulators (the hard and soft palate, the alveolar ridge, the teeth2) whose tactile proprioception is systematically carried out by sensors located in all the relevant organs; vocal actions whose perturbing effect in the surrounding atmospheric environment is immediately perceived by auditory means; and finally, vocal actions whose muscular counterparts are at least partially visible on any speaker’s face from the hearer’s position3. Any word can be defined in the terms of selected sensori-motor couplings on the side of the signifier, but on that of the signified, words refer to ‘objects’ and ‘events’ whose experience implies a whole range of highly heterogeneous sensations: a church is a place with a given visible structure, but also a place in which one will stand, sit and occasionally kneel and walk, have contact with stone and wooden elements, hear songs, psalms, readings and homilies, smell incense. In this example, none of these competing sensations is selected: the word church does not display any submorpheme. In the word spire, one selected visual impression attached to the object, that of pointedness, is selected by one submorpheme, sp, itself a sensori-motor complex in terms of vocal action.

19Because of this mapping, it is tempting to look for analogical matches between linguistic and non-linguistic experiences. This is the choice implicitly made by most submorphologists: Tournier, for instance, explicitly states that ideophonic elements are motivated by an unidentified synesthaesia (as opposed to onomatopœia, in which a linguistic sound mimics a natural sound without requiring any perceptual shift: cock-a-doodle-doo). Tournier does not suggest that submorphemes might well be just as unmotivated or de-motivated as fully-fledged words, and the effect of this bias is manifest in his presentation of English ideophonic elements: the table indicates (i) the marker, sp, (ii) the purported meaning, ‘pointed’, but nothing is said about (iii) the iconic connection, that is, what kind of senso-mimetic connection there should be between the two clusters of impression, the vocal action sp on the one side and the experience of a spire on the other – where does one feel, see or hear sp to involve pointedness? One might even add that st sounds and feels more adequate than sp for this kind of impression. In the same way, the connection between gl- and luminosity is widely acknowledged (Bolinger 1940, 1950, 1953, reprinted in Bolinger 1965), in English as well as in Semitic languages; but in present-day English, the kind of synesthesia that might link the impression left by the vocalization with that left by the object, is a mystery. So a reasonable attitude might consist in (a) investigating the existence of submorphemes as lexical-semantic classifiers in English and (b) cautiously hypothesizing remote diachronic motivations involving onomatopœia and then synesthesia. However, these connections might be far too distant to be traceable, and they do not explain why submorphology remains active in the current lexicon. Thus, leaving aside the question of motivation, ideophonic elements should be understood as sensophonic classifiers.

  • the impressive value of submorphemes is objectively difficult to determine and therefore highly exposed to the danger of subjectivism and unfalsifiability. In some cases there emerges a ‘natural consensus’ over semantic values: gl- is consistently associated with luminosity by authors, st- with fixity, sw- for pendulation (swing, sweep, swap, swim, swoon), so that there is no real space for controversy. In some cases, the consensus is motivated by a relatively obvious onomatopœic background: /skr/ as in screech, scratch, scrub) for a sound-generating disarticulated movement applied to a surface. In other cases, the consensus seems to be caused by analogous non-auditory sensations which should not be treated in terms of synesthesia either: wr- is said to underline the notion of torsion in wring, writhe, wriggle, write and many other words; in this series the possible analogy lies between the tactile experience of twisting a body or an object and the tactile proprioception of the movement of the tongue required for producing an apical /r/. But in most cases, the semantic unity is not especially clear, and nor are the sensori-motor connections, so that varying semantic classes may be proposed:

Fig. 2: The diverging designations of submorphemic semantic values






Centrifugation, projection


Spreading, “blooming”

Centrifugation under

agentive control


Standing position, fixity




Stativity under energetic

control: tension





Slight torsion

Accelerated pendulation

leading to the synthesis of







Region of the nose

Nasality (seen as a conceptual domain)



Rapid movement

Cutting (seen as a conceptual domain)

Rapid movement in relation to a surface

20This mismatch is motivated by three methodological protocols underlain by three different projects:

  • Tournier’s goal is to arrive at the heuristic proposal of a semantic classification on the basis of intuition. Introspection is the rule, and subjectivity is compensated by empirical investigation into English-speaking subjects’ own intuitions: according to Tournier, a majority of interrogated subjects will agree on the values ascribed to ideophonic elements, either immediately, or after some reflection. In Tournier’s approach, some words are implicitely regarded as central, prototypical representatives of the class: spear, spike and spindle for sp and pointedness. As a result, an ideophonic element’s value can usually be glossed by a term containing the very submorpheme in question, and this is illustrated by unilingual dictionaries’ tendency to define words using terms containing the same cluster.

  • Bottineau’s project is to account for the embodied nature of both the semantic experience of the signified and the articulatory experience of the signifier – in short, to bridge the gap between the two domains and to reduce a binaristic, symbolic conception to a functionally unified one.

  • Philps’s aim is to show that while the Saussurian sign is arbitrary by definition and by conception, certain submorphemic components of the sign (e.g. sn- in English ‘sn- words’ related to the conceptual domain of /nasality/) may have become arbitrary as convention and self-organizing, system-internal constraints took over from the essentially iconic conditions purported to underlie the emergence of the sign (semiogenetics). In particular, the dyadic {sn-/nasality} relation may, at some indeterminably early stage in the evolution of language, have resulted from the putative self-referential potential of the nasal sound n seen as a ‘core invariant’, in which case there would exist a causal link between the terms of this relation.

3. The experience-based semantics of submorphemic markers

21As a general principle, submorphemic elements do not underline homogeneous collections of actions and objects. Rather, they emphasize one salient, and therefore relevant, piece of multimodal sensori-motor experience through which the object can be apprehended; submorphemes relate concepts with keynote experiences through which they are captured:

  • /sp/: centrifugation, rapid rotation, ejection. This submarker is found for actions and objects which are seen as rotating things (spindle) or events (spin); it also appears in objects whose ordinary use implies a rotating movement on the part of the human agent: spoon, spear, spade. A spoon is “rotated” for the purpose of collecting food or throwing a stone (roman catapults); a spear is “rotated” when thrown away; a spade is “rotated” when manipulated.

  • /st/: stability, fixity. This submarker is found (i) in verbs expressing the absence of movement or change: stand, stay, still; (ii) verbs expressing the terminal access to a state of stability: stop, stoop. This terminal state may concern either a human agent, or the object he manipulates as an instrument: sting, stab, stick.

  • /sw/: pendulation. This marker is found in (i) verbs of human attitude or action involving a bodily oscillation (swoon, sway, swim), (ii) transitive verbs implying the use of an instrument or object with the corresponding movement (sweep, swap, switch), (iii) intransitive verbs referring to the alternating movements of non-human external objects (swing, swivel).

22As is illustrated by the preceding examples, one ideophone is applicable to the three conceptual domains spanning from the subject to the object:

Fig. 3: The phenomenological focus of submorphemic elements





speak, spit, spew

spoon, spear, spade

spindle, spawn


stand, stutter, stumble

stir, start, stop

(in transitive uses)

stick (the object), star, stub


swim, sway

switch, swap


23What the above table reveals is that the spatial locus of the property marked by the submorpheme on the subject-object scale is not taken into account. This is confirmed by hybrid cases in which an object is submorphologically labelled in the terms of the action implied for an implicit but necessary agent: a switch is an appliance that is meant to be switched by an agent; a Swatch (a brilliant neologism based on an ‘*s- mobile’ reminiscent of the Indo-European one) is a watch one will catch sight of by making a circular movement of the arm (a rotation of the sp class, indeed), and then coming back to the initial position. This reciprocal contamination of the subject’s and object’s dynamic properties is most visible in the case of conversions (to switch, a switch; to stitch, a stitch) and valency alternations (to stop: (semantically) reflexive, transitive). A combination of the two processes can cause one lexical item to cover the three levels:

Fig. 4: The co-variation of the submorphemic phenomenological focus and of the lexical categorization (word class and valency): the case of stop






To stop (doing something)


To stop someone


A bus stop

24The lexical submarker focuses the interpreter’s attention on one dynamic sensori-motor coupling that will ordinarily characterize the human subject’s interaction with the object (be it a thing, an event, an action). Whether this interaction is motoric (a spoon) or visual (to spin) is irrelevant; whether it primarily affects the subject (to spew), the object (a sparrow) or both (a spoon) is ignored: perception is a type of action that will involve sensori-motor couplings and the retrieval of multimodal knowledge of the experience of the object / event / action and the way in which both the subject and the object are reciprocally affected in the interaction. In short, the submarker gives substance to the dynamic experience co-determining the perceiving-acting subject and the object targeted in the environment. In this sense, ideophonic markers can be defined as phenomenological classifiers: they do not describe or symbolize the external, objective reality; instead, they provide the interpreter with a prediction of the way in which an object might behave (a spider), or be used (a spoon), or the sensori-motor way in which an event or action is likely to be experienced (to spew, to spawn). In other words, submarkers make a subjective comment on the detection of objective ‘things’: they indicate how experience is, in Varelian terms, enacted. In Varela’s theory of perception (Varela et al. 1993), enaction is the way in which the perception of the environment by living agents consists in mapping the sensation (a bottom-up input or signal) against the anticipation (a top-down counteracting output or signal) of what subsequent sensations, happenings and relevant actions might be (e.g., the anticipation of edibility, a pleasant smell and taste, and pleasure on seeing and hearing sizzling bangers or melting chocolate in a TV commercial).

25In the same way, the word sponge is a vocal action (articulating “sponge”) and perception (sensing oneself wording sponge through the tactile captors located in the tongue and all the active and passive articulators of the oral tract; hearing oneself or anybody voice the word), that is, a multimodal sensori-motor coupling whose effect is to activate “meaning”, a hierarchized network of “features” (as they are defined in the first section). According to the referentialist approach to language, this context is first and foremost “the object to which the word refers”, that is, the thing on which the word is supposed to focus the interpreter’s attention. As it happens, the word may very well be used in the absence of the corresponding object, if any; it has even been pointed out that new-borns and infants hear the French word cochon “pig” in comparisons like tu en mets partout comme un cochon “you’re spilling your food all over the place like a pig” long before they ever come across the animal, even in picture books.

26In developmental terms, this shows that the recorded meaning of the word is primarily the stereotypical discourse accompanying it in social situations involving prototypical actions and agents (in this case, eating and the family): “a pig is something which spills its food as I do”; and secondarily, the stereotypical association may be mapped against recurring non-linguistic impressions – in this case the “pig” itself, once encounters with the animal (in books and/or in real life) stabilize the connection and are recorded as a match (“this is a pig”). In one exceptional case, I heard an English mother use the words pig, spill and spread about her child, and she heard herself doing that, so that she eventually came up with the playful portmanteau word spig: “a pig that is likely to spread its food”. In so doing, she had embedded a submorphemic element, sp, into the lexical unit pig: she had enacted the “sp-prediction” for pigs and thus adjusted the label to the current relevance that motivated her applying the term to her own child.

27More generally, the submorphemic marker makes the same kind of culturally stabilized pragmatic comment on the notion reactivated by the lexical unit. This explains the randomness of the phenomenon:

  1. an object may well not be marked by the comment it obviously calls for: in English a broom is not an ‘sw- word’, on account of the synecdoche plant > object. In Spanish, escoba is marked by the /sk/ submorpheme ‘movement applied to a surface’ that is found in English words like scour, scrub etc.: broom might also have been an sk- word, and is indeed one in other languages. Book is kniha in Czech, knigu in Russian: an articulated object comprising two moving parts and a hinge like a knee, a knob, a knife, to knit, a knot etc.; book might have been a ‘kn-word’ in English, but is not.

  2. in one given language, the diverse communities that have taken part in the introduction of lexical strains may have passed diverging pragmatic comments on the same objects and activities, so that in the present-day language, different “visions” will cohabit. For instance, to write ‘refers’ to an activity which is characterized by wr- as a specific movement, the experience of torsion (cf. wring, wrist, wrestle, writhe, wriggle, wrath, wrought, wrap, wreck, wretch, wry…). But in the Latin strain, scribere has left script, inscribe, scribble, scrabble, écrire in French, escribir in Spanish, schreiben in German, all of which involve the scr- submorpheme: a friction applied to a surface by an agent. This marker may have resulted from two merging strains, sk+r (surface + agentivity: scrub, scroll, scratch) and the scr- onomatopœia (screech, shrill). Clearly, two competing pragmatic outlooks have shed diverging submorphemic light on the same notion, and the socio-linguistic signature of the comments is engrained in the lexical distribution: wr- is reserved for ordinary experience in daily life (write) while scr- is reserved for abstract applications and technological jargon.

28These elements indicate that submorphemic markers should be treated as inbuilt pragmatic cues developed in the course of the linguistic tradition and culture attached to a given speaking community: they do not indicate the way in which an object determines an individual subject’s perception or action, but the random way in which a community has constructively focused on a selected cue at the expense of potentially available alternative ones. The notion of ‘cue’ raises the question of the cognitive relevance of submorphemes for individual subjects and how a submorphemic competence or sensitiveness may be formed in language acquisition.

4. Do speakers actually develop a submorphemic know-how?

29Submorphemic units represent a serious phenomenological challenge: the linguist detects semantic consistency which emergences from lists of words that he has created on his own; this semantic response is motivated by the very presentation of the object of analysis, and the risk of circularity has to be circumvented. At least two possibilities can be envisaged:

  • either the detection of submorphemic semantic consistency occurs if and only if the interpreting subject is confronted with lexical strings that make the isomorphisms detectable and favour the construction of analogical connections. In this case it has to be shown the ordinary speakers are given the opportunity to experience submorphemic resemblances together and are likely to compare them and derive semantic links. Otherwise the linguist’s phenomenological co-determination of the object and his own self remains idiosyncratic and submorphology is not experienced by the speaking community. Even so, one would have to explain why the lexicon tends to self-organize diachronically in a way that is apt to foster this kind of semantic mirage.

  • or the detection of this semantic unity is obtained directly by the speakers, without the help of experiential stimuli. The theoretical cost of this hypothesis is extremely high as it requires either the phonosymbolic determination of the form / meaning connections, or, even worse in our view, the innate programming of those links in universal grammar, making them independent from experience.

  • A middle-of-the-road position consists in proposing that form-meaning associations have to be initiated by the experience of privileged discourse contexts, fostered by linguistic practices and scaffolded by the speaking community before the subject is able to go on incorporating additional lexical units into pre-existing ensembles.

4.1. Submorphemic elements in language acquisition

30The detection of ideophones as semantically relevant submorphemes requires at least some of the following features: (i) the number of lexical units whose stressed syllable bears semantically ‘colourful’ submorphemic clusters has to be sufficiently high in relation to ‘uncolourful’ ones in order to be detectable and noted. And (ii) the possibility for the subject to detect isomorphisms from which semantic analogies can be drawn has to be established. It is easy for the linguist to associate sleep with slumber on the basis of the sl- connection in the list s/he has created, but in the subject’s spontaneous experience of language it seems unlikely that the sleep / slumber relation may emerge unless the two words are regularly encountered in a shared environment in which this very connection is emphasized. An empirical survey suggests that this is the case in nursery rhymes and lullabies:

31(i) selected words are repeated a number of times, hence facilitating the remembering process, and the child is given the opportunity to practise the sensori-motor dynamics of articulation:

A swarm of bees in May
Is worth a load of hay;
A swarm of bees in June
Is worth a silver spoon;
A swarm of bees in July
Is not worth a fly.

32(ii) Some lullabies and nursery rimes are loaded with ideophonic clusters:

Twinkle, Twinkle, Little Star
Twinkle, twinkle, little star,
How I wonder what you are!
Up above the world so high,
Like a diamond in the
Twinkle, twinkle, little star,
How I wonder what you are!

33(iii) Rhymes consist in emphasizing a final constant and distinguishing an initial variable; to create a rhyme is to cause the orientee to detect an analogical link between two previously unrelated lexical items.

Humpty Dumpty
Humpty Dumpty sat on a wall,
Humpty Dumpty had a great f
All the King’s horses, and all the King’s men
Cannot put Humpty Dumpty together again.

34And indeed one of the requirements for a rhyme to be accepted as such is that it should never associate two identical suffixes as done by Vincent Voiture (c1630), a practice parodied in Molière’s Tartuffe and denounced in Voltaire’s Rhétorique:

Estampe et crampe vrayment,
Riment admirablement.

35The following lullaby is remarkable:

Angels To Watch Over You
When you go to bed at night
And you turn off all your
In the dark and out of
There’s angels to watch over you.

36This is not to suggest that -ight forms a submorphemic element in its own right, but that this lullaby forms the kind of relevant candidate from which a pattern is potentially derivable: night = “vision is obscured”, light = “vision is favoured” and sight = “vision” (unspecific / unmarked); so in this local network -ight = “vision”, n- = negative, l- = positive and s- = neutral. For this local network to be extrapolated and stabilized, it has to be confirmed by other instances like Then you show your little light, / Twinkle, twinkle, all the night in Twinkle, twinkle little star; also, it should not be counterpoised by an overwhelming number of contrary examples. In this respect, the case of -ight is dubious, as it may be confirmed by bright and white (so long as the spelling is ignored) but not by right and fight; initial n- is regularly connected with negation (and in Slavic languages the d-/n- contrast underlies both the semantic day / night pair and the grammatical yes / no pair: Russian den / not’, da / niet). So the kind of environment formed by this poem is not sufficient to determine the systematic and universal construction of submorphemic analogies framing all learners’ lexical systems, but it may well participate in the emergence of submorphemic connections in the learner’s personal experience encounters other similar networks.

37(iv) Words bearing the same submorpheme are frequently found together:

Itsy-Bitsy Spider
The itsy-bitsy spider
Climbed up the water
Down came the rain
And washed the
spider out
Out came the sun
And dried up all the rain
And the itsy-bitsy
Climbed up the spout again

38(v) In some cases, pairs of words (or several words) which are manifestly connected both semantically and ideophonically, regularly appear together in a number of lyrics, to such an extent that they tend to form a stereotype, a culturally established discourse connection or link between lexical units in common association of ideas:

Golden slumbers
Golden slumbers kiss your eyes,
Smiles awake you when you rise;
Sleep, pretty baby, do not cry,
And l will sing you a lullaby.

All Through the Night
Sleep my child and peace attend thee,
All through the night
Guardian angels God will send thee,
All through the night
Soft the drowsy hours are creeping,
Hill and dale in
slumber sleeping
I my loved ones’ watch am keeping,
All through the night
Angels watching, e’er around thee,
All through the night
slumber close surround thee,
All through the night
Soft the drowsy hours are creeping,
Hill and dale in
slumber sleeping
I my loved ones’ watch am keeping,
All through the night

39(vi) Some submorphemes which refer to sensori-motor processes are usually scaffolded by gestures executed by the mother to illustrate the meaning of the song:

This old man
This old man, he played one;
He played
knick-knack on my thumb.
With a
knick-knack, paddy whack,
Give a dog a bone;
This old man came rolling home.

The process is repeated ten times with as many different body locations.

40One should not, of course, exaggerate the power of nursery rhymes and lullabies. What really matters is the extent to which they are intuitively used as linguistic standard-setting models by parenting adults teaching their offspring to recognize words, understand their meaning and practise their articulation. Children are actually taught to rely on submorphemic analogies to form semantic networks and learn associative stereotypes, and the question is whether they do cross a critical threshold beyond which they become autonomous analogists and start incorporating new isolated lexical items into pre-existing submorphemic ensembles without the stimulus of rhyme, paronomasis and so on.

4.2. The popular awareness of submorphemic dynamics

41According to Tournier, ideophonic elements are no longer productive, insofar as English-speaking communities do not create new clusters any more. Tournier does not say that existing clusters are fossils which have fallen under the level of intuitive detectability. Indeed, traces of neologisms motivated by submorphology are evidenced, both in academic and popular literature: in Dicken’s A Christmas Carol, the protagonist’s name is Scrooge, in reference to the notion of scratching the soil for food, and hence making the name suggestive of a rodent’s or vulture’s. The name of one of the characters in the film Trainspotting is Spud, and, as mentioned earlier, a famous trademark has produced Swatches. In 1983, Douglas Adams and John Lloyd published a small, humorous dictionary entitled The Meaning of Liff. It consists in a list of place names located in English-speaking countries (most of them in the British Isles); these names are taken to be nouns and given a definition:

BANFF (adj.): Pertaining to, or descriptive of, that kind of facial expression which is impossible to achieve except when having a passport photograph taken.

GLASGOW (n.): The feeling of infinite sadness engendered when walking through a place filled with happy people fifteen years younger than yourself.

SKETTY (n.): Apparently self-propelled little dance a beer glass performs in its own puddle.

42Some connections are inspired by direct lexical plays on words: Hastings is “things said on the spur of the moment to explain to someone who comes into a room unexpectedly precisely what it is you are doing”. But in many cases submorphemic clusters are instrumental in inspiring a lexical interpretation for the toponym: Blean (“a measure of luminosity”), Glossop, Sneem, Stebbing (“an unconcealable erection”), Skegness, Spofforth, Sproston Green, Swanage, Wrabness, Writtle, etc., hundreds of toponyms are selected for their colourful sonority – and the potential correlations borne by the submorphemic elements they contain, frequently echoed in the definitions. Obviously the authors were receptive to the sense-making effect of the submorphemic hypo-sub-actions and trusted their potential readers to share this semantic experience.

  • 4 The ARTFL Project, University of Chicago: (...)

43Folk etymology is another domain in which submorphology will strive. Here is an excerpt from an online academic medical dictionary (which borrowed its information from various sources, including Webster’s 1913 edition, accessible online4):

SCALL : a scurf or scabby disease, esp. of the scalp.
SCALD < excaldare: to run off like a scalded cat ; a scald, a boil
SCALL < ON skalli “bald head”
A scald is a burn caused by very hot liquid or steam.
Variant of skald.
Variant of

  • 5

44A skald is normally a bard while a scald is a burn on the skin. Hundreds of occurrences of this scald / skald confusion can be googled up, and it seems quite obvious that we are dealing with competing analogies: scald / scab on the one side, skald / skin on the other. In the same way scall, from Old Norse skalli ‘bald head’, is mistakenly considered as a “variant of scall”, and Skall is the name of a French body artist whose project is presented in the following texts5:

Visual artist before anything, Skall considers his body as a raw material to produce art. In his performances, he invites the spectator to assist in a transformation, more specifically, his transformation. The transformation that takes place is not only physical but is also of psychological nature. He does not play, he is his proper creation. Each character evolves in function with the physical environment or sound. If we would have seen this mutant character being born, we would also assist in his death; the spectator therefore sees an act of shamanism or an act of mysticism...referring to Greek or Indian mythology. Skall performed different projects for the Annual Rodeo Performance in 2002 and 2003 which takes place during the month of July.

Plasticien avant tout, Skall considèrent son corps comme la matière première d’une œuvre plastique. Dans ses performances, il invite le spectateur à assister à une transformation, sa propre transformation aussi bien physique que psychologique : il ne joue pas, il est sa propre création. Chaque personnage évolue en fonction de l’environnement visuel ou sonore et si on a vu naître ce caractère mutant, on assiste aussi à sa mort ; le spectateur peut alors y voir un acte de chamanisme ou un acte de mysticisme se référant à la mythologie grèque ou indienne. Il a participé avec differents projets aux Rodéo Performances 2002 et 2003 qui ont lieu le mois de juillet à camac.

45Although the English translation was obviously not written by a native English speaker, the choice of the very name ‘Skall’ does evidence the awareness of the submorphemic motoric correlations which are explicitely mentioned in the presentation.

46Some speakers are fully aware that there may exist a mismatch between etymology and the way in which lexical units are reorganized in present-day English. Here are two suggestions for the origin of the word slang:

The origin of the word “slang” is unknown, but it may come from the term “thieves” language” where the possessive “s” from thieves” and the “lang” from language are run together to make the word slang.
So now we are all “on the same page,” as it were, with regard to our definition of
slang, let us at this time turn to the origin of this noun. Etymonline informs us that slang derives in all likelihood from “a Scandinavian source, cf. Norwegian slengenamn “nickname,” slengja kjeften “to abuse with words,” literally “to sling the jaw,” related to Old Norwegian slyngva ‘to sling’.”
However, both Etymonline and the OED agree that these etymological suppositions are just that, and that "the ultimate source [of the word] is not apparent.”
At any rate, it appears that even the word slang itself is colorful. In other words, I sling slang. I slung slang with far-flung, fun-in-the-sun abandon. And so on...

47Beyond the contradiction between the two proposals, the second one clearly states that knowing the origin of the word may not be cognitively relevant in the determination of its present-day colour, which is not explicitly defined, but is exclusively based on the submorphemic common denominator: colorful is here the intuitive formulation of what we have called the semantic effect of the hypo-sub-action. Here is a discussion about the word snobe, found in a forum6:

Can somebody answer this stupid riddle for me please?
If all snubes were snobs, and all snobes are snoobs, then why dont the snoobs get along with the snobs if they both think the same way as the snobes? but not all snobs are friends with snoobs because the snobes never liked the snorps, by the way the snorps are not related to the snoobs, nor the snobes, but the snobs liked them. If the snobes dont like the snorps, and the snorps didnt like the snobs, then nobody liked anybodys family, until the snobies came along...the snobies were a special family, and started hanging out with the snorps, who didnt like the snobs, because all they wanted to do was snoop. So the snobies thought they were being sort of snooty, and told them to shut up. But then Sally stepped in, and told the snobies not to talk to the snobs, because the snorps dont like the snobes anymore than the snobs like the snobies. So who is talking to who now?

48This of course is an extreme example in which the writer has deliberately played with sounds, but it must be added that submorphemically motivated turns of phrases and idioms do tend to emerge in everyday language, like to go on a spending spree, scant and scarce (scarce < excerptus and scant < ON skammr “short”):

Although it is one of the driest deserts in the world, with scant and scarce rainfall, when it does rain, the wildflowers, like these lupins and colorful yellow flowers, bloom for a short time.

49In my corpus, it appears that the order of the adjectives is overwhelmingly i) scarce and ii) scant in predicative position and the other way round in attributive function:

Most of the land is arid and desert; rainfall is scarce, vegetation scant, and very little of the land is suitable for agriculture.

Alternative energy sources are scarce and scant. Only one institution reported their use, a pilot photovoltaic installation.

50In this combination, sk- plays the role of the unvarying common denominator in relation to which an -arce / -ant variable is articulated. In context, scarce takes on a heuristic value: it provides an objective measurement of the quantity; and in relation to it, scant is hermeneutic: it adds an evaluative undertone and passes a modal judgment on the underlying quantification. This sequence can be analyzed as an operation / result sequence (the discovery and the evaluation of the quantity). In attributive constructions this schemata is preconstructed and lexicalized, which is marked by the inversion (cf. to surge up > upsurge, to hunt whales > whale-hunting, be scarce and scant > scant and scarce N). The submorphemic marker fixes a trajectory that can be improvised and run forward or retrieved and run backwards.

5. Conclusion

51Investigation into the cognitive relevance and popular availability of submorphemic markers is incipient. The latter can be defined as the markers of the sensori-motor couplings through which objects and events are culturally assumed to be apprehended: vocal hypo-sub-actions with a characterizable contribution in the distributed sense-making process. Considering the evidence, there is good reason to believe that they cannot be merely explained by the lexicon’s apparent tendency to self-organize as an autonomous system disconnected from the intersubjective interplay and subjective experience; and they must be taken into account to describe the subject’s lexical-semantic linguistic know-how.

52On the basis of the problematic we have outlined above, systematic corpus and field research is required to quantify and formalize the distribution of the submorphemic experience in language acquisition and verbal practices in families and schooling systems, but also in suburban social communities, with all the social variation that this implies. Since the submorphemic competence emerges from specific verbal practices in differentiated social environments, one should not expect there to be a fully unified treatment of the phenomenon. However, there remains the possibility that these practices share common values.

Top of page


Anscombre, Jean-Claude, 2001). « Le rôle du lexique dans la théorie des stéréotypes », Langages, 142, 57-76.

bar-Lev Zev, 2008. « Key consonants in Arabic » (pre-published version), Journal of Arabic and Islamic Studies 6 (2005–2006).

Bateson, Gregory, 1979. Mind and Nature, a necessary unity, Bantam.

Benveniste, Émile, 1939. « Nature du signe linguistique », Acta Linguistica, I, Copenhagen, reprinted in (1966) Problèmes de linguistique générale, 1, Gallimard, 49-55.

Bergen, Benjamin K., 2004. “The psychological reality of phonæsthemes”, Language, 80, 2, 290-311.

Bohas, Georges, 2006. “The organization of the lexicon in Arabic and other Semitic languages”, in Boudelaa, S. (ed.), Perspectives on Arabic Linguistics XVI, Papers from the sixteenth annual symposium on Arabic linguistics, Cambridge, March, 2002, John Amsterdam/Philadelphia, John Benjamins, 1–37.

Bolinger, Dwight L., 1965. Forms of English: Accent, Morpheme, Order, Isamu Abe and Tetsuya Kanekiyo (eds), Cambridge, Harvard University Press, Tokyo, Hokuou.

Bottineau, Didier, 2002. « Les cognèmes de l’anglais : principes théoriques », Lowe, R. (dir.), en collaboration avec Pattee, J. et Tremblay, R., Le système des parties du discours, Sémantique et syntaxe, Actes du IXe colloque de l’Association internationale de psychomécanique du langage, Québec, Les Presses de l’Université Laval, 423-437.

Bottineau, Didier, 2003a. « Les cognèmes de l’anglais et autres langues », Ouattara, A. (éd.), Parcours énonciatifs et parcours interprétatifs, Théories et applications, Actes du Colloque de Tromsø organisé par le Département de Français de l’Université, 26-28 octobre 2000, Gap, Ophrys, 185-201.

Bottineau, Didier, 2003. « Iconicité, théorie du signe et typologie des langues », Monneret, Philippe (dir.), Cahiers de linguistique analogique, no 1 – Juin 2003, Le mot comme signe et comme image : lieux et enjeux de l’iconicité linguistique, Association Bourguignonne d’Etudes Linguistiques et Littéraires (ABELL), Dijon, 209-228.

Bottineau, Didier, 2004. « Le problème de la négation et sa solution dans la langue anglaise : le cognème N », Delmas, Claude & Roux, Louis, La contradiction en anglais, C.I.E.R.E.C. Travaux 116, Publications de l’Université de Saint-Etienne, 27-53.

Bottineau, Didier, Didier, 2006. « Le cognème <M>, marque linguistique de la présence de l’auteur dans les grammèmes anglais », Bank, David (éd.), Les marqueurs linguistiques de la présence de l’auteur, Paris, L’Harmattan, 143-164.

Bottineau, Didier, 2007. « Language and enaction », Stewart, J., Gapenne, O. & Di Paolo, E. (eds), Enaction: towards a new paradigm for cognitive science, MIT, in press.

Bybee, Joan, 1985. Morphology, Amsterdam/Philadelphia, John Benjamins.

Coursil, Jacques, 2000. La fonction muette du langage, Guadeloupe, Ibis Rouge Éditions, Presses Universitaires Créoles.

Culioli, Antoine, [1981] 1990. « Sur le concept de notion », BULAG no8, repris dans Pour une linguistique de l’énonciation, t.1 : Opérations et représentations, Gap, Ophrys.

Danon-Boileau, Laurent, 1991. « De quelques préjugés relatifs à l’usage des notions de motivation et d’iconicité », Faits de langues, 1, Motivation et iconicité, Paris, PUF, 79-87.

Davis, Joseph, Gorup, Radmila J. & Stern, Nancy (Eds.), 2006. Advances in Functional Linguistics, Columbia School beyond its origins, Studies in Functional and Structural Linguistics 48, Amsterdam/Philadelphia, John Benjamins.

Dennett, Daniel Clement, 1991. Consciousness Explained. Boston, Little, Brown and Co.

Diver, William, 1979. “Phonology as human behavior,” in D. Aaronson and R. Rieber (eds.), Psycholinguistic research: implications and applications. Hillsdale, N.Y., Lawrence Erlbaum Assoc., 161-182.

Douay, Catherine, 2000. Eléments pour une théorie de l’interlocution, Un autre regard sur la grammaire anglaise, Rennes, Presses Universitaires de Rennes.

Duchet, Jean-Louis, 1990. « Arbitraire et motivation dans le lexique et la morphologie de l’anglais », Roux, L. (ed), L’organisation du sens, domaine anglais, Recueil en l’honneur de Jean Lavédrine, Travaux du CIEREC, LXVIII, 57-66.

Edelman, Gerald M., 1989. The Remembered Present: a Biological Theory of Consciousness, New York, Basic Books.

Engler, Rudolf, 1962. « Théorie et critique d’un principe saussurien. L’arbitraire du signe », Cahiers Ferdinand de Saussure, 19, 5-66.

Engler, Rudolf, 1964. « Compléments à l’arbitraire », Cahiers Ferdinand de Saussure, 21, 25-32.

Feldman, Jerome, & Narayanan, Srini, 2004. “Embodied meaning in a neural theory of language”. Brain and Language, 89, 385-392.

Firth, John R., 1930. Speech, London, Ernest Benn.

Fischer, Olga & Nänny, Max (Eds.), 2001. The Motivated Sign : Iconicity in Language and Literature 2, Amsterdam/Philadelphia, John Benjamins, 249-276.

Gallese, Vittorio, & Lakoff, George, 2005. The brain’s concepts: the role of the sensory-motor system in conceptual knowledge. Cognitive Neuropsychology, 22 (3/4), 455-479.

Gardiner, Alan H., [1932] 1951. The Theory of Speech and Language, Oxford, Clarendon Press.

Genette, Gérard, 1976. Mimologiques, Paris, Seuil.

Givon, Talmy, 1994. “On the Co-evolution of Language, Cognition and Neurology”. Paper for 10th meeting of Language Origins Society, Berkeley.

Grammont, Maurice, 1933. Traité de phonétique, Paris, Delagrave, 377-424.

Gregory, Richard Langton, 1976. Concepts and Mechanisms of Perception. London, Duckworth.

Guiraud, Pierre, 1986. Structures étymologiques du lexique français, Paris, Payot.

Heine, Bernd, 1997. Cognitive Foundations of Grammar, New York, Oxford University Press.

Householder, Fred W., 1946. « On the Problem of Sound and Meaning, an English Phonestheme », World 2, 83 sq.

Jacob, François, 1970. La logique du vivant, Paris, Gallimard.

Jakobson, Roman & Waugh, Linda R., [1979] 1980. La charpente phonique du langage, Paris, Minuit.

Keller, Pascal-Henri, 2006. Le dialogue du corps et de l’esprit, Paris, Odile Jacob.

Kemp, J. Alan (1972), John Wallis’s Grammar of the English Language, London, Longman.

Kravchenko, Alexander V., 2004. “Essential properties of language from the point of view of autopoiesis”,

Lakoff, George & Johnson, Mark, 1999. Philosophy In The Flesh: the Embodied Mind and its Challenge to Western Thought, New York, Basic Books.

Langacker, Ronald W., 1987. Foundations of Cognitive Grammar. Stanford, Stanford University Press.

Leroi-Gourhan, André, 1964. Le geste et la parole. I. Technique et langage. II. La mémoire et les rythmes, Paris, Albin Michel.

Lieberman, Philip, 1991. Uniquely Human: The Evolution of Speech, Thought, and Selfless Behavior. Cambridge (Mass.), Harvard University Press.

Love, Nigel, 2007. “Are languages digital codes?”, Language Sciences 29, 690-709.

Luquet, Gilles, 1996. “Arbitrariness and Non-Arbitrariness in the Historical Evolution of Language” in Tollis, F. (ed.), The Psychomechanics of Language and Guillaumism = LynX (Valence et Philadelphie), 5, 154-157.

Macchi, Yves, 1986. « Du rôle du signifiant dans la genèse du sens énonciatif », Langages, 82, 67-82.

Macchi, Yves, 1991. « Halte aux manipulations sémiques : de la métaphore chimique en sémantique », Modèles Linguistiques, 13/1, 25 101-116.

Macneilage, Peter F., 1998. « The frame/content theory of evolution of speech production », Behavioral and Brain Sciences, 21, 499-546.

Marchand, Hans, 1959. « Phonetic Symbolism in English Word-Formation », Indogermanische Forschungen, 64, 146-68, 256-77.

Markel, Norman N. & Hamp, Eric P., 1960-1. « Connotative Meanings of Certain Phoneme Sequences », Studies in Linguistics, 15, 47-61.

Maturana, Humberto R., 1978.Biology of language: The epistemology of reality”, in Miller, G., and E. Lenneberg (eds.), Psychology and Biology of Language and Thought: Essays in Honor of Eric Lenneberg, New York, Academic Press, 1978, 27-64.

Maturana, Humberto R., & Varela, Francisco J., 1980. Autopoiesis and Cognition: the Realization of the Living. Reidel, Dordrecht.

Maturana, Humberto R., & Varela, Francisco J., 1987. The Tree of Knowledge: the Biological Roots of Human Understanding, Shambhala, Boston, London.

McNeill, David, 2005. Gesture and Thought, Chicago & London, University of Chicago Press.

Molho, Maurice, 1958. “Acerca de la evolución del grupo -skl- -sl- en romance”, Archivo de filología aragonesa, 8-9, p. 63-71.

Molho, Maurice, 1980. « Verbe et personne. De l’arbitraire du signe », Cahiers de linguistique hispanique médiévale, 5, 5-23.

Molho, Maurice, 1986. « Grammaire analogique, grammaire du signifiant », Langages, 82, Le signifiant, 41-51.

Molho, Maurice, 1988. « L’hypothèse du « formant » : sur la constitution du signifiant : esp. un/no », Grammaire et histoire de la grammaire, Hommage à la mémoire de Jean Stéfanini, recueil d’études rassemblées par Blanche-Benveniste, C., Chervel, A. & Gross, M., Publications de l’Université de Provence, 291-303.

Nemo, François, 2004. “Morphemes and Lexemes versus Morphemes or Lexemes”, in Sergio Scalise (ed), Proceedings of the 4th Mediteranean Morphology Meeting (MMM4). on Morphology and Linguistic Typology. (Catania, Sicile, 21-23 Septembre 2003).

Newman, Stanley S., 1933. « Further Experiments in Phonetic Symbolism », American Journal of Psychology, 45, 53-75.

Oudeyer, Pierre-Yves, 2006. Self-organization in the Evolution of Speech, Oxford, OUP.

Peirce, Charles Sanders, [1894]. “What is a Sign”, In: Houser; Kloesel (eds). The Essential Peirce. Selected Philosophical Writings. Vol 2. (1893-1913), Indiana University Press.

Philps, Dennis, 1997. « A la recherche du sens perdu : <sn->, du marqueur au mythe », Anglophonia / Sigma 2, Toulouse, Presses Universitaires du Mirail, 209-38.

Philps, Dennis, 2003a. « L’invariance sub-lexicale et le marqueur <sk-> », Anglophonia / Sigma 14, Toulouse, Presses Universitaires du Mirail,177-193.

Philps, Dennis, 2003b. « S- incrémentiel et régénération submorphémique en anglais », Bulletin de la Société de Linguistique de Paris vol. XCVIII, fasc. 1, 163-196.

Putnam, Hilary, 1975. “The meaning of ‘meaning’”, Mind, Language and Reality, Cambridge, Cambridge University Press, 215-271.

Reid, Wallis Hoch, Otheguy, Ricardo & Stern, Nancy (Eds.), 2002. Signal, Meaning, and Message, Perspectives on sign-based linguistics, Amsterdam/Philadelphia, John Benjamins.

Robert, Stéphane (éd.), 2003. Perspectives synchroniques sur la grammaticalisation, Louvain-Paris, Peeters,

Robert, Stéphane & Chapouthier, Georges, 2006. “The Mosaic of Language”, Marges linguistiques – Issue n°11, May 2006 - M.L.M.S. Publisher, 160-166. - 13250 Saint-Chamas (France)

Salzen, Eric A., 2006. “From Calls to Words : How ethology can bridge the divide”, Marges linguistiques - Numéro 11, Mai 2006 - M.L.M.S. éditeur, 200-217, - 13250 Saint-Chamas (France)

Schogt, Henry, 1992. « Onomatopées, mots expressifs et la métaphore en phonétique », Mélanges Pierre Léon, Martin, P. (dir), Toronto, Mélodie, Canadian Scholars’ Press et Paris, Didier Erudition, 463-71.

Shanon, Benny, 1993. The Representational and The Presentational: An Essay on Cognition and the Study of Mind. London, Harvester-Wheatsheaf.

Stewart, John, 1996. “Cognition = Life: Implications for higher-level cognition”. Behavioural Processes 35: 311-326.

Studdert-Kennedy, Michael, & Goldstein, Louis, 2003. Launching language: the gestural origin of discrete infinity. In M. H. Christiansen & S. Kirby (Eds.), Language evolution (pp. 235-254). Oxford, Oxford University Press.

Talmy, Leonard, 2000. Toward a Cognitive Semantics, volumes I and II, Cambridge (Mass.), MIT Press.

Tobin, Yishai, 1997. Phonology as Human Behavior: Theoretical Implications and Clinical Applications, Durham & London, Duke University Press.

Tollis, Francis, 1998. « Le couplage analytique des morphèmes en /un/- et en -/l/-, vu de l’espagnol », in Leeman, Danielle, Boone, Annie et alii, Du percevoir au dire. Hommage à André Joly, Paris, L’Harmattan, 132-145.

Tournier, Jean, 1985. Introduction descriptive à la lexicogénétique de l’anglais contemporain, Paris - Genève, Champion – Slatkine.

Toussaint, Maurice, 1983. Contre l’arbitraire du signe, Paris, Didier.

Varela Francisco, Thompson Evan T., & Rosch Eleanor (1993). The Embodied Mind: Cognitive Science and Human Experience, Cambridge (Mass.), MIT Press.

Vygotsky, Lev S., 1962. Thought and Language. E. Kaufamnn & G.Vakar, eds & trans., Cambridge (Mass.), MIT Press.

Wallis, John, 1653. Grammatica linguæ anglicanæ. R.C. Alston (Ed.), Reprint 142 (1969). Menston, The Scolar Press.

Waugh, Linda, 1993. « Lexique : iconicité diagrammatique », Faits de langues, 1, Motivation et iconicité, PUF, 227-34.

Whorf, Benjamin Lee, 1956. Language, Thought, and Reality: Selected Writings of Benjamin Lee Whorf, Ed. J.B. Carroll, Cambridge (Mass.), MIT Press.

Wilson, Margaret, 2002. “Six views of embodied cognition”. Psychological Bulletin and Review, 9(4), 625-636.

Ziemke, Tom, 2002. “What’s that thing called embodiment?” in Proceedings of the 25th Annual Meeting of the Cognitive Science Society, Lawrence Erlbaum, 1305-1310.

Top of page


1 This metaphorical, somewhat Californian-sounding terminology is only meant to emphasize the collaboration of vocal and mental dynamics, which are not understood as separate processes. The elaboration of an adequate and, hopefully, more elegant terminological apparatus is still under way, but we feel an urgent need to override the reifying effect of conventional and static terminology.

2 The teeth are used as passive articulators against the tongue, but as active articulators against the lips in the case of labio-dental fricatives. In English the lips are used as active articulators (bi-labials and labio-dentals) but not as passive ones (there exists no such thing as apico-(bi)labial phonemes).

3 In Stanley Kubrick’s film 2001: A Space Odyssey, the computer Karl manages to ‘interpret’ the astronauts’ conversation by just watching the movements of their lips as they talk. This is of course impossible, as not all articulators are visible, unless one postulates that the action of internal articulators has secondary effects on external ones which can be detected and plotted. However, the visual perception of muscular action, along with bodily gestures and eye contact, is instrumental in sense-making processes in non-vocal sign languages.

4 The ARTFL Project, University of Chicago:


6 Currently

Top of page


Electronic reference

Didier Bottineau, « The submorphemic conjecture in English: towards a distributed model of the cognitive dynamics of submorphemes », Lexis [Online], 2 | 2008, Online since 10 November 2008, connection on 29 April 2017. URL : ; DOI : 10.4000/lexis.688

Top of page

About the author

Didier Bottineau

CNRS (MoDyCo – Université Paris 10-Nanterre)

Top of page
  • Logo Centre d’études linguistiques
  • Logo Université Jean Moulin – Lyon 3