Research, Society and Development, v. 15, n. 10, e38151051805, 2026 (CC BY 4.0) | ISSN 2525-3409 | DOI: http://dx.doi.org/10.33448/rsd-v15i10.51805 From process to form: Foundations for a cognitive-praxic refoundation of epistemology. Concept, language, communication and understanding before representation Do processo à forma: Fundamentos para uma refundação cognitivo-práxica da epistemologia. Conceito, linguagem, comunicação e entendimento antes da representação Del proceso a la forma: Fundamentos para una refundación cognitivo-práxica de la epistemología. Concepto, lenguaje, comunicación y entendimiento antes de la representación Received: 10/03/2026 | Accepted: 10/07/2026 | Published: 10/09/2026 Giovanna Azevedo Rodrigues ORCID: https://orcid.org/0009-0003-9020-9368 ASMIGO Psychiatric Hospital, Brazil E-mail: giovannagirodrigues@hotmail.com Marcelo Caixeta ORCID: https://orcid.org/0000-0003-2357-6558 ASMIGO Psychiatric Hospital, Brazil Universidade Federal de Goiás, Brazil E-mail: cfm11041963@ufg.br Abstract This theoretical article examines the Kantian theory of the concept, as reconstructed by Robert Hanna, and proposes from the standpoint of Cognitive Psychiatry an inversion of its explanatory order. The objective is to show that concept, object, unity, representation, sign, rule and understanding are late results of cognition and cannot serve as its original conditions. The method is a qualitative conceptual essay supported by a narrative review of classical sources and of studies published between 2021 and 2026. Hanna’s formulation is decomposed into six theses, each confronted with a genetic reconstruction that begins in the lived and praxic processuality in which organism and world are already causally coupled. The results describe a sequence that runs from praxis through interindividual coupling and synchronisation to objectification and formalisation, and only then reaches signification, concept, representation and propositional language. To be told to another person, lived experience has to be chewed into communicable units, which requires a model of the listener’s mind and a model of one’s own; both derive from the mother, the first verbal accompanist of the infant’s praxes, whose accompaniment, once internalised, helps to constitute the secondary self. The vocalisation “ai” in pain illustrates functional communication prior to semiosis: the sign arises when a stabilised praxis becomes decoupled from its original situation and able to evoke it. Recent developmental studies converge with this ordering. It is concluded that form does not found process; process progressively founds form, which calls for a genetic epistemology of pre-objectal, pre-formal and pre-semiotic processualities. Keywords: Knowledge; Concept Formation; Language Development; Communication; Psychiatry. Resumo Este artigo teórico examina a teoria kantiana do conceito, tal como reconstruída por Robert Hanna, e propõe, a partir da Psiquiatria Cognitiva, a inversão de sua ordem explicativa. O objetivo é mostrar que conceito, objeto, unidade, representação, signo, regra e entendimento são resultados tardios da cognição e não podem servir como suas condições originárias. O método é o ensaio conceitual qualitativo, apoiado em revisão narrativa de fontes clássicas e de estudos publicados entre 2021 e 2026. A formulação de Hanna é decomposta em seis teses, cada uma confrontada com uma reconstrução genética que parte da processualidade vivencial-práxica na qual organismo e mundo já estão causalmente acoplados. Os resultados descrevem uma sequência que vai da praxia ao acoplamento interindividual e à sincronização, daí à objetificação e à formalização, e só então alcança significação, conceito, representação e linguagem proposicional. Para ser dita a outro, a vivência precisa ser mastigada em unidades comunicáveis, o que exige um modelo da mente do interlocutor e um modelo da própria mente; ambos derivam da mãe, primeira acompanhante verbal das praxias do bebê, cujo acompanhamento, internalizado, ajuda a constituir o eu secundário. A vocalização “ai” diante da dor ilustra uma comunicação funcional anterior à semiose: o signo surge quando uma praxia estabilizada se desacopla da situação original e passa a evocá-la. Estudos recentes do desenvolvimento convergem com essa ordenação. Conclui-se que a 1 Research, Society and Development, v. 15, n. 10, e38151051805, 2026 (CC BY 4.0) | ISSN 2525-3409 | DOI: http://dx.doi.org/10.33448/rsd-v15i10.51805 forma não funda o processo; o processo funda progressivamente a forma, o que requer uma epistemologia genética das processualidades pré-objetais, pré-formais e pré-semióticas. Palavras-chave: Conhecimento; Formação de Conceito; Desenvolvimento da Linguagem; Comunicação; Psiquiatria. Resumen Este artículo examina la teoría kantiana del concepto, según la reconstruye Robert Hanna, y propone, desde la Psiquiatría Cognitiva, invertir su orden explicativo. El objetivo es mostrar que concepto, objeto, unidad, representación, signo, regla y entendimiento son resultados tardíos de la cognición y no pueden ser sus condiciones originarias. El método es el ensayo conceptual cualitativo, apoyado en revisión narrativa de fuentes clásicas y de estudios de 2021 a 2026. La formulación de Hanna se descompone en seis tesis, confrontadas con una reconstrucción genética que parte de la procesualidad vivencial-práxica en la que organismo y mundo ya están causalmente acoplados. Los resultados describen una secuencia que va de la praxia al acoplamiento interindividual y la sincronización, luego a la objetivación y la formalización, y solo entonces alcanza significación, concepto, representación y lenguaje proposicional. Para ser dicha a otro, la vivencia debe ser masticada en unidades comunicables, lo que exige un modelo de la mente del interlocutor y otro de la propia; ambos derivan de la madre, primera acompañante verbal de las praxias del lactante, cuyo acompañamiento, internalizado, ayuda a constituir el yo secundario. La vocalización “ay” ante el dolor ilustra una comunicación funcional anterior a la semiosis: el signo surge cuando una praxia estabilizada se desacopla de la situación original y pasa a evocarla. Estudios recientes del desarrollo convergen con esta ordenación. Se concluye que la forma no funda el proceso; el proceso funda progresivamente la forma, lo que requiere una epistemología genética de las procesualidades preobjetales, preformales y presemióticas. Palabras clave: Conocimiento; Formación de Concepto; Desarrollo del Lenguaje; Comunicación; Psiquiatría. 1. Introduction Modern epistemology usually opens its analysis with items that are already formed: the representation, the concept, the rule, the judgement, the proposition. Kant’s theory of the concept is the most carefully built version of this starting point, and Robert Hanna’s reconstruction states it in a single sentence: “every concept is the result of a specific generative mental operation, or rule-governed act, of synthesis derived from the understanding, applied to a finite set of characteristics, each of which is originally abstracted from the perceptual representations indicative of objects” (Hanna, 2001/2004, p. 193, our translation from the Brazilian edition). The sentence concentrates several presuppositions of transcendental epistemology: the precedence of the mental, the generative function of the understanding, the existence of rules prior to the concept, the synthesis of a manifold, the representation of objects that are already constituted and the abstraction of characteristics from those representations. They are Kant’s own commitments. In the Critique of Pure Reason concepts rest on functions, understood as the unity of the act of ordering different representations under a common one, and the understanding is described as the faculty of concepts, of rules and of judging (Kant, 1781/1998, A68–69/B93–94, A126). In the Jäsche Logic the form of a concept is produced by three logical acts of the understanding, namely comparison, reflection and abstraction (Kant, 1800/1992, §6). A different picture has been assembled over the last century by the study of development and of embodied cognition. Piaget (1936) described a sensorimotor intelligence that precedes representation and proposed that epistemology should be genetic (Piaget, 1970). Vygotsky (1934/1986) traced inner speech to the internalisation of social speech. Trevarthen (1979) documented a primary intersubjectivity in the first months of life, and Bruner (1983) showed how the caregiver’s routines frame the child’s entry into language. Merleau-Ponty (1945), Gibson (1979), Varela et al. (1991) and Clark (1997), each in a different vocabulary, placed perception and cognition in the dealings of a body with its surroundings before any inner picture of the world. Studies from the last five years add detail to this picture: infants learn that their vocalisations change what adults do months before they have words (Elmlinger et al., 2023), turn-taking in the proto-conversation of mother and infant is accompanied by interpersonal neural synchrony (Nguyen et al., 2023), mothers’ verbs follow what the infant is doing at that moment (West et al., 2022), and in thirteen languages caregivers simplify their speech when they answer a child’s immature vocalisations (Elmlinger et al., 2025). These results are seldom brought to bear on the epistemological question raised by Hanna’s sentence, namely what 2 Research, Society and Development, v. 15, n. 10, e38151051805, 2026 (CC BY 4.0) | ISSN 2525-3409 | DOI: http://dx.doi.org/10.33448/rsd-v15i10.51805 the concept is the result of. The thesis presented here, formulated within Cognitive Psychiatry, proposes to invert almost the whole of that explanatory order. The concept does not begin in a generative mental operation. It begins in a lived and praxic processuality in which organism and world are already causally coupled. Objectification comes later, as do formalisation, signification and the semiotic function. The understanding, in the sense of a faculty that handles representations and applies rules, is later still. In the most primitive genesis of communication there is no need to postulate concepts, mental objects, representations or signs: a bodily praxis can become coupled to a vocal praxis, the vocalisation of one individual can reorganise the praxis of another, and two processualities can synchronise before either of them symbolically “means” the other. The elementary example is the vocalisation “ai”, the Portuguese counterpart of “ouch”, in the face of pain, which shows that functional communication is possible before semiosis: at first the listener does not have to relate a word to a representation of “pain” and can simply synchronise his or her action with the processuality of the speaker. The genetic architecture proposed runs as follows: processuality → lived experience and praxis → coupling → interindividual synchronisation → recurrence → vocal accompaniment → internalisation of the accompaniment → selfreflection → objectification → formalisation → signification → concept → representation → propositional language. The middle of this sequence is occupied by communication: the mother’s verbal accompaniment of the infant’s praxes, its internalisation as secondary self, and the operations by which lived experience is chewed for a listener. If concept, object, unity, representation, sign, rule and understanding are late results, an epistemology that starts from them risks turning products of cognition into its original conditions. The consequences are therefore epistemological and not only psychological. The aim of this article is to examine, thesis by thesis, the Kantian formulation reconstructed by Hanna, to set out the foundations of a cognitive-praxic refoundation of epistemology in which form is the outcome of process, and to verify how far this ordering converges with literature published in the last five years. The proposal does not intend to abolish the formal. It relocates it genetically: form ceases to be foundation and becomes processual sedimentation. 2. Methodology This is a qualitative study of a theoretical and conceptual nature (Pereira et al., 2018), conducted as an argumentative essay supported by a narrative review of the literature. The object of analysis is the Kantian theory of the concept in the formulation given by Hanna (2001/2004), read against the primary texts (Kant, 1781/1998, 1800/1992). The analysis followed three steps. First, the formulation was decomposed into its constituent theses. Second, each thesis was confronted with a genetic reconstruction developed within Cognitive Psychiatry, which asks what must already have taken place in the organism and between organisms for the item named by the thesis (mental operation, rule, synthesis, understanding, characteristic, representation of an object) to be available. Third, the resulting ordering was compared with classical works on development, embodiment and language and with recent empirical and theoretical studies. For the third step, a non-systematic search was carried out in October 2026 in PubMed/PubMed Central, PhilPapers and publisher and institutional repository pages, with combinations of the terms concept formation, Kant, enactivism, embodied cognition, interpersonal synchrony, proto-conversation, infant vocal development, caregiver speech, interjections, pain vocalisation, inner speech and causal learning. Peer-reviewed articles published between 2021 and 2026, in English, that bore directly on one of the steps of the proposed sequence were included; texts without identifiable authorship or without peer review were excluded. The search was updated during the revision of the manuscript, with priority given to studies published from 2024 onwards. Twenty recent studies were retained, eleven of them from 2024 to 2026, and are cited at the points of the argument on which they bear. Because the study is theoretical and involves no human participants or identifiable data, submission to a research 3 Research, Society and Development, v. 15, n. 10, e38151051805, 2026 (CC BY 4.0) | ISSN 2525-3409 | DOI: http://dx.doi.org/10.33448/rsd-v15i10.51805 ethics committee was not required. As in any narrative review, the selection is not exhaustive. It serves to situate and to test the argument, not to estimate effects. 3. Results and Discussion 3.1 The Kantian text as a point of rupture Hanna’s sentence can be decomposed into six theses: (i) the concept results from a mental operation; (ii) this operation is generative; (iii) it is specific; (iv) it is governed by rules; (v) it consists in a synthesis derived from the understanding; (vi) it acts upon characteristics previously abstracted from perceptual representations of objects. The criticism developed here holds that all six begin the analysis too late. They are not necessarily false as a retrospective description of an adult cognition that is already formalised. They become problematic when they are used to explain the genesis of the concept itself. The distinction matters because the Jäsche Logic presents comparison, reflection and abstraction as the acts by which concepts are generated as to their form (Kant, 1800/1992), and the question of whether comparison already requires the concept it is supposed to produce is still debated among Kant scholars (Erlewein, 2026; Wang, 2021). Erlewein (2026) presents it as a trilemma: the account of concept formation seems bound to move in a circle, to regress or to break off. The difficulty noted inside the Kantian tradition is the one the genetic reading takes as its point of departure: the items invoked to explain the concept are themselves of conceptual grain. The exposition follows the genetic sequence and not the order of the sentence. Section 3.2 deals with the first three theses. Sections 3.3 to 3.7 describe what the sentence leaves out: the communicative operations by which experience acquires form, the mother’s accompaniment in which they originate and the secondary self that carries them out. Sections 3.8 and 3.9 return to theses (iv) and (v), sections 3.10 and 3.11 examine language and the sign, and section 3.12 takes up thesis (vi). The last sections assemble the architecture, its consequences and its theses. 3.2 Before the mental operation: lived experience and praxis The first disagreement concerns the expression “mental operation”. Before an object concept is formed there is no need to suppose a mental operation already organised over representations. What exists at first is an organism in a situation. It acts, is struck, changes posture, moves, avoids, approaches, vocalises, perceives and responds. If an animal chases me, before I construct conceptually that “a lion chased me” there is visual movement, a roar, a threat, autonomic change, flight, a change of direction, fear and bodily orientation. The starting point is lived and praxic, and it is neither formal nor representational. The mind does not begin by inspecting an internal representation of an object, because the organism is already inside a processuality. This starting point is the one described by the classical literature on sensorimotor intelligence and embodied perception (Clark, 1997; Gibson, 1979; Merleau-Ponty, 1945; Piaget, 1936; Varela et al., 1991), and recent work pushes it further back. Ciaunica et al. (2021) observe that the human organism first develops inside another human body, so that its earliest regulation is a co-embodiment and a co-homeostasis and the minimal self is relational from the outset. Di Paolo (2021) argues that the enactive approach requires bodies and minds to be treated as open and unfinished processes of becoming, and not as entities given in advance. Both positions describe organism and world as coupled before anything is represented. The expression “generative mental operation” raises a second difficulty. It may suggest an active mind that deliberately produces a concept. In the theory presented here a large part of the initial process is automatic, reflex, neurophysiological, nondeliberate and non-conscious. The subject does not think “I shall now synthesise my perceptions in order to generate the concept of lion”; that would be an intellectualist reconstruction of the process. Conceptualisation emerges from processes that long precede deliberate thought. 4 Research, Society and Development, v. 15, n. 10, e38151051805, 2026 (CC BY 4.0) | ISSN 2525-3409 | DOI: http://dx.doi.org/10.33448/rsd-v15i10.51805 Nor is the operation “specific”. There is no special operation whose task is to generate the concept of lion. The mechanism is broader: to live through, to act, to be affected, to communicate, to cut out, to objectify. Specificity belongs to the lived material, and the mechanism is general. There is no faculty for LION and another for STONE. There is a general cognitivepraxic architecture. 3.3 Communicative volition and the need to chew experience Although the process is largely automatic, there is a moment at which the will gains importance: I want to say something to someone. This communicative wish changes the processuality deeply. I now have to look at what happened to me, select elements, summarise, cut, order and adjust my expression to the person I am addressing. Formalisation therefore does not arise only from a supposed internal need of the understanding. It is strongly driven by a relational need, that of making my experience accessible to another person. The volitional element is specific only as to its communicative destination. What happens cognitively afterwards, the selection, retrieval, association and ordering, depends largely on automatisms that are already sedimented. The process is a mixture of communicative volition and cognitive-neurophysiological automatisms, and at no point is it a voluntary construction supervised in full. A lived experience is too rich to be transmitted whole. In the chase everything happened at once: the shape that moved, the sound, the tightening of the body, the ground under the feet, the fear. None of it came divided. Speech can only say one thing after another. Vygotsky (1934/1986) gave the difference an image: thought is a cloud and speech is the rain of words that falls from it, so that what is present all at once in thought has to be unfolded in succession when it is spoken. The same holds, with greater reason, for lived experience that has not yet become thought. To tell it I have to turn it into blocks, cuts, units and sequences. It has to be, metaphorically, chewed. Chewing is made of the operations already named. To cut is to break the flow at certain points and detach pieces from it: the animal, the roar, the running. To select is to leave most of what was lived unsaid and to keep what can carry the episode for someone who was not there. To summarise is to let one piece stand for many, so that “it chased me” covers a long stretch of turns, stumbles and glances over the shoulder. To order is to place the pieces in a line, with a before and an after, although in the event several of them were simultaneous. To adjust is to choose, for each piece, the size and the wording that this particular listener can take in. None of these operations is found in the event. The chase had no joints at which it asked to be cut. The joints are introduced by the need to tell it. This chewing has two simultaneous poles, the event as lived and the capacity of the interlocutor. Food is chewed for a mouth, and experience is chewed for a listener. The same episode is cut in one way for a child, in another for a hunter who knows the animal, and in yet another for someone who was at my side and only needs to hear “it came back”. I do not simply describe “the object”. I produce a communicable version of my experience with it, and the version changes with the person who receives it. The role of language changes accordingly. Language does not serve only to express concepts that are ready beforehand. It takes part in the objectification of experience. When I have to explain that “it was brown”, “it had a mane”, “it came fast”, “it roared” and “it tried to reach me”, I am converting a processual experience into communicable units. Communication forces cutting, objectification and formalisation. To communicate is not simply to transmit form; communicating helps to produce form. Borghi (2022) reaches a compatible conclusion from the study of concepts: all concepts are grounded in sensorimotor processes, and the more abstract ones depend to a greater degree on linguistic and social interaction for their acquisition and use, so that other people are part of how the concept comes to exist. A more recent review of this line of work brings together the evidence that social interaction takes part both in the formation and in the use of abstract concepts (Borghi et al., 2025). Merleau-Ponty 5 Research, Society and Development, v. 15, n. 10, e38151051805, 2026 (CC BY 4.0) | ISSN 2525-3409 | DOI: http://dx.doi.org/10.33448/rsd-v15i10.51805 (1945) had made the point from the side of the speaker: speech does not translate a thought that is already complete, it accomplishes it, and the speaker finds out what there was to say in the saying. What Merleau-Ponty says of thought holds here for the object. Before the telling there was a chase. After the telling there is a lion that is brown, maned, fast and loud. The units into which the experience was chewed for the listener stay with the speaker. They are the first draft of the characteristics that a later analysis will find “in” the object, and of the object itself as that which has them. 3.4 To speak to the other I have to model the other’s mind When I communicate I cannot simply pour out my experience. I have to ask implicitly what my interlocutor knows, what he or she can recognise, which terms are familiar, how much needs explaining and what can be left out. The operation resembles what is now called theory of mind: modelling cognitively what the other is able to understand. The model does real work in the formation of what is said. If I judge that my listener has never seen a lion, the word “lion” is useless and I have to go down to “a big animal, like a cat, with hair around its head”. If I judge that the listener knows the animal well, the single word is enough and the detail would be noise. The grain at which experience is cut is therefore fixed from outside, by an estimate of another person’s knowledge. The estimate is revised as the exchange goes on. A frown, a question or a silence tells me that the piece was too big or too small, and I chew again. Communication is shaped by the expectation about the interlocutor’s cognitive capacity and corrected by the interlocutor’s response. Tomasello (1999) describes the developmental ground of this operation. Towards the end of the first year the child comes to treat other persons as intentional agents like itself, enters scenes of joint attention with them and learns words inside those scenes. A word learned in this way carries a perspective. The same animal can be called lion, animal, beast or danger, and to choose one of these is to choose how the listener is asked to take the situation. The linguistic form is perspectival because it was made for someone. The adjustment to the other begins even earlier, before there is anything to say: by five months infants expect their vocalisations to change what the adult does and vocalise more when the expected response fails (Elmlinger et al., 2023). The model of the listener’s knowledge that the adult speaker exercises has its first form in this practical anticipation of the listener’s conduct. This has a consequence for the theory of the concept. A form produced for one listener is still tied to that listener. But I tell the episode again, to other people who know other things, and each telling demands a new cut. What survives the succession of listeners, what can be said to any of them and recognised by all, is the part of the experience that no longer depends on who is listening. The stable, re-identifiable and shareable pole that the tradition calls the object is reached along this path. It is what remains constant across interlocutors, and the modelling of other minds is one of the operations that extract it. 3.5 A theory of mind of myself A second operation is less discussed. To explain to someone what happened to me I must make my own lived experience accessible to myself. I have to reconstruct what I felt, what I did, what I saw, what frightened me, what came before and what came after. Communication thus involves a kind of theory of mind applied to oneself, an autotheory of mind. The operation is needed because lived experience is not transparent to the one who lives it. While I was running I was not observing my fear. I was afraid, and the fear was in the legs. The processuality was fully there, and it was not there as something looked at. To tell it I have to go back over it from another position, that of someone who asks what happened, and from that position my own experience has to be modelled much as the mind of another person is modelled: by signs, by reconstruction, with gaps and with error. I conclude that I was frightened from what I did. I settle that the roar came before the 6 Research, Society and Development, v. 15, n. 10, e38151051805, 2026 (CC BY 4.0) | ISSN 2525-3409 | DOI: http://dx.doi.org/10.33448/rsd-v15i10.51805 running because the account requires an order. I have to make my own processuality partly observable to me. It is not enough to know the other. I must be able to become, in part, an object of my own observation. Two things follow. The first is that telling requires a doubling. There is the one who lived and the one who reports, and the second treats the first almost as a third person. The questions I put to my own experience are the questions a listener would put: what was it, what did it do, what did you do, what happened then. The autotheory of mind is the theory of mind turned round. It borrows the interlocutor’s place and looks at oneself from there. The second is that the two models work together. I cut my experience according to what the listener can take in, and I find out what there is to cut by questioning myself as the listener would. Self-knowledge of this kind is a product of the effort to be understood. The literature on inner speech describes this turning round. Vygotsky (1934/1986) observed that speech for oneself is abbreviated, almost reduced to predicates, because the one who speaks and the one who listens share the whole context, and that speech for another has to be unfolded in proportion to what the other lacks. When I prepare to tell what happened to me I pass from the first regime to the second: I treat myself as a listener who needs the unfolded version. Fernyhough and Borghi (2023) review evidence that inner speech ranges from condensed to expanded forms, keeps the give and take of dialogue and gives the subject a distance from itself that serves reasoning and self-regulation. Borghi (2022) adds that with the more abstract concepts the subject monitors its inner knowledge for longer and then turns to other people to settle the meaning. Tomasello (1999) places the origin of this reflexive capacity in the child’s internalisation of the perspective that adults take on the child’s own thinking. In all these accounts the access of the subject to itself passes through the form of an exchange with someone else. 3.6 The mother as first verbal accompanist of praxes The adult capacities described in the three preceding sections, to chew experience, to model the listener and to model oneself, do not appear suddenly. In the present hypothesis they have an ontogenetic history, and the history begins with someone else doing all three on the child’s behalf. From very early the mother, or whoever occupies her place in the care of the child, accompanies the infant’s gestures, postures, affects, discomforts, pleasures, actions, apparent intentions and movements, and accompanies these praxes with words, intonations, rhythms, interjections, namings and punctuations. The baby lives, and she accompanies the living with vocalisation. This repeated association creates a fundamental matrix. She is the first verbal accompanist of the child’s praxes. The scene is ordinary. The baby stretches towards something and the mother says “you want it… there, you got it”. The baby startles and she says “oh, that was loud”. It pushes the spoon away: “no more, all done”. It falls, and before the crying has taken shape she has already said “ai, that hurt”. In each case a vocal form is laid alongside a praxis while the praxis is under way or has just ended. The baby did not ask for a description and is not yet able to use one. What it receives is a second contour, made of sound, that runs parallel to its own movement and affect, with the same onset, the same peak and the same end. Three features of this accompaniment matter for the theory. It is contingent in time: the vocalisation arrives with the praxis, and it is this coincidence, repeated, that couples the two. It is rhythmic and intonational before it is lexical: the rise of the voice follows the rise of the effort, the falling “there” marks the completion, and the infant can follow the contour long before it can follow the words. And it punctuates: by beginning and ending where the action begins and ends, the voice marks off one stretch of the flow from the next. The first cuts in the child’s experience are made from outside, by a voice that says where something started and where it finished. Observation gives this description empirical content. Recording mothers and infants at home, West et al. (2022) found that the verbs mothers used corresponded to the actions infants were performing in real time, so that the word arrived while the infant’s own body was carrying out what the word named. The coincidence is older than words. Trevarthen (1979) had filmed, 7 Research, Society and Development, v. 15, n. 10, e38151051805, 2026 (CC BY 4.0) | ISSN 2525-3409 | DOI: http://dx.doi.org/10.33448/rsd-v15i10.51805 in infants of two months, exchanges in which the mother’s voice and face answer the infant’s movements and expressions with matching timing and intensity, and called this primary intersubjectivity; in dyads with infants of four to six months, the amount of vocal turn-taking is related to interpersonal neural synchrony and to the size of the child’s later vocabulary (Nguyen et al., 2023). The coincidence is also repeated. Bruner (1983) described the recurrent formats, such as peekaboo, give-and-take games and looking at picture books, in which the same praxis meets the same vocal accompaniment day after day, and noted that mothers treat the infant’s movements from the start as if they were meant. A longitudinal study of 104 dyads, observed at four, six, nine and twelve months, found that the context of play shapes these early exchanges and that book-sharing favours vocal turn-taking (Laudańska et al., 2025). Much of what is said in these exchanges consists of interjections, which Dingemanse (2024) shows to be working tools of interaction. Wittgenstein (1953, §244) had seen the essential point in the case of pain: the child hurts itself and cries, and the adults talk to it and teach it exclamations and, later, sentences. What the mother does in these scenes is what the adult speaker will later do alone. She chews. She breaks the infant’s continuous activity into stretches that receive a word, she picks out of everything the infant is doing the part she names, and she offers it in a size the infant can take in: a single word, repeated, with an exaggerated contour. The metaphor is exact here. As with food, the first chewing is done by another mouth. The reduction in size has been measured: in 1,586 transcripts from thirteen languages, caregivers’ speech was simpler when it answered a child’s immature vocalisations than at other moments, with fewer different words, shorter utterances and more utterances of a single word (Elmlinger et al., 2025). She also models the infant’s mind. The apparent intentions she puts into words (“you want it”) are a theory of what the child is after, held by her before the child can hold any theory of itself. And she adjusts continually to what the infant shows of having followed. The infant’s experience thus reaches the infant already partly cut, selected and named by a person who has a model of it. The pattern has a precedent in the body. Ciaunica et al. (2021) describe the earliest regulation of the organism as cohomeostasis, carried out with and through another body before it is carried out alone. The verbal accompaniment of praxes repeats the arrangement on another plane: the infant’s experience is punctuated by another before it is punctuated by the infant. Later this accompaniment can be internalised. The subject comes to do with itself what at first someone else did with it. The external function, “someone accompanies my experience with language”, gradually becomes “I accompany my own experience with language”. The intermediate step can be heard. The young child playing alone says aloud what it is doing, in the mother’s words and often in her intonation. Vygotsky (1934/1986) studied this speech for oneself and noted that at first it comes with the action or at its end, as a commentary, that it then moves towards the beginning of the action and starts to direct it, and that it finally goes silent as inner speech. The child has taken over the role of accompanist. The possibility of narrating oneself would thus have an interpersonal genealogy: an external caregiver accompanies praxis with vocalisation, the association recurs, it is internalised, and the result is an internal accompaniment of one’s own praxes. To have the mother inside oneself is the condition for telling oneself what is happening. There is an evident proximity here with Vygotsky’s idea of the internalisation of social speech, which the current literature on inner speech broadly supports (Fernyhough & Borghi, 2023). Vygotsky helps to understand the internalisation of language. The present theory intends to go further back and asks about the earlier phase in which vocalisation does not yet need to be fully signifying language. Before social speech that is semantically constituted there would be vocal-motor praxic couplings that are not yet properly semiotic. The infant receives the mother’s first accompaniments as sound that goes with its action, and not yet as statements about it. The coupling is established at this level, and only because it is already established can the same vocal forms later be taken as meaning what they had accompanied. In the vocabulary of Cognitive Psychiatry, the internalised accompaniment is related to the constitution of the secondary self, to which the next section is devoted. 8 Research, Society and Development, v. 15, n. 10, e38151051805, 2026 (CC BY 4.0) | ISSN 2525-3409 | DOI: http://dx.doi.org/10.33448/rsd-v15i10.51805 3.7 The secondary self as a communicative acquisition Cognitive Psychiatry distinguishes two levels of the self. The primary self is the organism that feels, acts, orients itself, wants, persists, remembers, expects, avoids and seeks without having to say “I am doing this”. It simply does. It is lived before it is thematised, and it is the self of section 3.2, the one who runs from the lion. The secondary self is the level at which the organism reacts to its own reactions. It perceives, accompanies, describes and reorganises the activity of the primary self. It is the self that looks at itself acting, and it is what makes possible deliberate thought, self-observation, explicit planning, the narration of oneself and thought about thought. Nothing in the primary self demands the second level. An organism can deal with the world with great precision and never turn back on its own dealing. The nervous system provides the possibility, since a reaction can be registered and become the input of another reaction. The possibility, however, does not settle from what position the first reaction will be looked at, nor with what instrument it will be held. The hypothesis of this article is that the position is first occupied by someone else. Before the infant reacts to its own reactions, the mother reacts to them. The infant stretches and she answers the stretching; it startles and she answers the startle. Each of her accompaniments is a reaction to a reaction, and it has the very functions that will later define the secondary self: it perceives the infant’s praxis, accompanies it, describes it and, by soothing, encouraging or stopping, reorganises it. In the first months the circuit of the secondary self therefore already exists, but it runs through two organisms. The primary activity is in the child and the return upon it is in the mother. It is the arrangement already met in bodily regulation (Ciaunica et al., 2021) and in the exchanges of the second month (Trevarthen, 1979): a function distributed between two organisms before it belongs to one. The infant registers this return early, as the anticipation described in section 3.4 shows (Elmlinger et al., 2023). The secondary self is formed when this return changes place. At first there is organism and praxis. Then the other reacts to me, names my actions and gives my experience back to me in words. I internalise this perspective, and a layer arises in which I can see myself as the other saw me. Three moments can be distinguished in the passage. In the first the return is external: it is the mother who says what I am doing. In the second the return is already mine, but it is still spoken aloud and still in her voice: it is the speech for oneself of the child at play, described in section 3.6 (Vygotsky, 1934/1986). In the third the return has become silent. It is what the adult knows as inner speech, and as the plain fact of noticing what one is doing. Recent developmental theory also places a level of this kind late. In the agency-based model of Tomasello (2024), a second-order tier that monitors and controls the child’s own executive processes emerges only after three to four years of age, and the flexibility of thought that results is traced to the child’s efforts to coordinate different social perspectives in activities carried out with others. What is internalised is a position more than a stock of words: the position of someone who watches me act and says what I am doing (Tomasello, 1999). This is why the phrase “to see myself as the other saw me” should be taken literally. The secondary self looks at the primary self from outside, and the outside it looks from is the place the mother occupied. Two features of adult self-observation are explained in this way. It typically has the form of an exchange, with someone who asks and someone who answers, and the literature on inner speech describes this dialogic form and the distance from oneself that it provides (Fernyhough & Borghi, 2023). And it cuts experience at the joints at which experience was first cut by another voice: I notice in myself the wanting, the hurting, the being afraid and the having finished that were once noticed in me and named. This does not make the secondary self a purely social product. The capacity of a nervous system to register its own reactions and to react to them is a biological condition, and a return upon one’s own activity that does not use words is conceivable. Adults in fact differ widely in how much inner speech they report, from almost constant to almost none (Nedergaard & Lupyan, 2024). What the communicative history supplies is the place from which the primary self is looked at and the vocal forms with which what is seen is cut and held. The internalised perspective of the other is therefore one constitutive element of 9 Research, Society and Development, v. 15, n. 10, e38151051805, 2026 (CC BY 4.0) | ISSN 2525-3409 | DOI: http://dx.doi.org/10.33448/rsd-v15i10.51805 the secondary self, and the secondary self is, for that reason, also a communicative acquisition. The secondary self is the agent of the operations described in sections 3.3 to 3.5. To chew experience for a listener, someone has to turn back on what was lived and cut it. To hold an autotheory of mind, someone has to take his or her own processuality as something to be looked at. That someone is the secondary self, and the questions it puts to experience (what was that, what did you want, did it hurt) are the questions that were first put to the child, with the answers supplied, by a person who was watching it live. Self-observation is already socially formed. When as an adult I ask “what happened to me?”, I am not performing a purely individual and original operation. I am using a capacity that was built relationally. In a certain sense I carry the interlocutor inside me, and the communicative function has become reflexive. Two consequences for the Kantian picture follow. The first concerns apperception. Kant places at the summit of the understanding an “I think” that must be able to accompany all my representations (Kant, 1781/1998, B131). The verb is the one used here for the mother. In the genetic reading, the accompaniment that Kant states as a formal condition has a history: before an “I think” could accompany my representations, a “you want it”, a “you fell” and a “that hurt”, said by someone else, accompanied my praxes. The unity that the reflective I brings to experience is the late and internal form of a company that was first external. The second consequence concerns the subject of knowledge. The subject that classical epistemology places at the start, the one who has representations and judges them, is the secondary self. If it is formed as described, the subject pole of the relation between subject and object is as late as the object pole, and both are made in the same history: the world is objectified by being told to a listener, and I am objectified to myself by taking the listener’s place. This gives the early relationship a deep epistemological role. The mother enters before the formal concept. Besides teaching names, she helps the organism to accompany, cut out, give rhythm to and punctuate its own experience, and this external punctuation, once it has become internal, is the reflective self with which the philosophical tradition begins its analysis. 3.8 Regularity before rule With the communicative operations and their ontogenesis in place, the remaining terms of Hanna’s sentence can be taken in turn. The expression “rule-governed act” supposes a cognitive architecture in which rules already exist and direct the formation of the concept. This follows from the role Kant assigns to the understanding as the faculty of rules. The present theory contests that priority. Before the formal rule there is processual regularity. There is something like a rule at the origin, though not in Kant’s sense. It consists of regularities that are far more primitive: something happens and then produces a consequence; what has occurred does not simply un-occur; time does not run back; an action alters a situation; processes accumulate transformations; certain organisations grow in complexity; certain causalities have a direction. At first these are regularities of lived processuality, and only later rules of the understanding. The true rule is praxic before it is logical. The organism learns in its own doing that action leads to consequence. It does not need to formulate a principle of causality. It does, undergoes, anticipates and corrects. Regularity is incorporated pragmatically before it is formalised. The order is: regularity of the world → praxis → incorporation → expectation → rule that can later be made explicit. It is not: transcendental rule → organisation of experience. Once a regularity is stabilised it can be stated as a rule, which makes the rule itself a late product and inverts Kant once more. Developmental research on causal cognition describes the same order. In their review, Goddu and Gopnik (2024) place the origin of causal understanding in the child’s learning to control its own movements and to act effectively on the environment, followed by targeted intervention, and they treat depersonalised and decontextualised causal representations, that is, objective and general ones, as what distinguishes mature human causal understanding. Explicit causal rules come at the end of a history of doing (Piaget, 1936). 10 Research, Society and Development, v. 15, n. 10, e38151051805, 2026 (CC BY 4.0) | ISSN 2525-3409 | DOI: http://dx.doi.org/10.33448/rsd-v15i10.51805 The hypothesis goes beyond simple empiricism. Besides saying that causality is learned by repetition, it holds that the organism itself is born of a universe that is already temporal, causal, directional, evolutionary and integrative. Biology does not create these processualities. It arises within them, and this is why the organism can incorporate them. This is the background of the idea, developed within Cognitive Psychiatry, of a processual form of intuition that is broader than space and time alone. 3.9 Synthesis, unity and understanding The next term is “synthesis”. To synthesise one must suppose elements A, B and C that are afterwards brought together. That division may itself be later. Before it there may simply be a processual flow. The separation into colour, movement, sound, threat and object is already the result of later cuts. At the beginning, then, there may be no manifold plus synthesis. There is a processuality from which manifold and unity will later be extracted. Synthesis and unity can be retrospective artefacts. Once LION has been formed we can reconstruct it as mane plus tail plus roar plus four legs, and it then seems that the mind synthesised these characteristics. We may, however, be analysing in retrospect a product that is already formalised. Synthesis can be a later logical description and not the original genetic operation. The same holds for unity. Kant repeatedly needs unity: unification, synthetic unity, unity of apperception (whose genetic counterpart was indicated in section 3.7), unity of the concept. If we do not start from a formal manifold, we do not need to start from a formal unity that gathers it. Unity can emerge afterwards. Integrative processes produce unity; unity is not what makes integration possible. This is a relevant criticism of the Kantian architecture, in which formalism creates the problem that unity is then called on to solve. If I start from separate elements, I have to explain how they come to be joined, and then synthesis, unity and a unifying I appear. If the point of origin is processual and has not yet been cut up, the problem may arise only because the theory formalised too early. Unity would be in part the solution to a problem produced by the formalist premise. Hanna says that the synthesis is “derived from the understanding”. Cognitively this raises a difficulty. The Kantian understanding is the faculty of concepts, the faculty of rules and the faculty of judging (Kant, 1781/1998). These are functional descriptions. They do not explain genetically how the faculty emerges, how its operations are constituted or why it takes those forms. The risk is that the name of the function is turned into the explanation of the function. Who synthesises? The understanding. And what is the understanding? That which synthesises according to rules. An explanatory circle becomes possible. The circularity that commentators discuss within Kant’s account of empirical concept formation, in which the comparison of representations appears to need the very concept it should yield (Erlewein, 2026; Wang, 2021), is a local form of this general problem. Conceptual activity can precede the formally articulated understanding. Primitive communication is again the simplest example. A baby can vocalise, respond, synchronise, anticipate and adjust, and there is no need to attribute to it a fully formed Kantian understanding. There is communicative and proto-conceptual activity before the understanding in the formal and reflective sense. If “understanding” names the capacity to recognise concepts, operate rules, judge and combine, it may itself be a late level built on more elementary processes, a result of sedimentations, and it should not be used without explanation as the origin of concepts. A metaphor sums this up: transcendental epistemology begins on the upper floor. Kant describes with mastery what is found there: concept, judgement, rule, synthesis, unity. The present theory asks how the building reached that height, and holds that its construction cannot be explained with the materials of the upper floor alone. 3.10 The cry of pain: communication before signification The interjection “ai”, announced in the Introduction and heard in the mother’s voice in section 3.6, is decisive for the argument. Someone steps on a thorn. There is pain, withdrawal, contraction and a vocalisation: “ai!”. The theoretical error would 11 Research, Society and Development, v. 15, n. 10, e38151051805, 2026 (CC BY 4.0) | ISSN 2525-3409 | DOI: http://dx.doi.org/10.33448/rsd-v15i10.51805 be to write at once that “ai” means PAIN. That presupposes signifier, signified, representation and semiotic function. None of this is needed. We can start only with an algesic-motor praxis coupled to a vocal praxis. One praxis can be coupled to another without representing it. This is one of the central points of the theory of language proposed here. The vocalisation does not need to represent the pain. It can simply accompany it. The two processes become coupled, and at first nothing requires a meaning. In the passage already cited, Wittgenstein (1953, §244) suggests that words for sensations are tied to the primitive and natural expressions of sensation and come to be used in their place, so that the verbal expression of pain replaces the cry and does not describe it. The listener does not need to decode either. Another subject hears “ai!” and may look, interrupt what he or she was doing, approach, draw back or help. The traditional description would be: heard a sign, interpreted “pain”, responded. A more economical description is available: the vocalisation of the other reorganises my praxis. This is interindividual synchronisation. The possibility breaks a very old association, that of language with sign and of sign with meaning. The thesis is that there can be functional communication before signification. The communication in question is coordination of conduct, still far from mature propositional communication. Before semiotics there is synchronisation: a speaker in process A produces vocalisation V, and a listener modifies process B. If this recurs, A, V and B become a relatively stable configuration. No symbol is needed at the start. This is a strong hypothesis: language may have a pre-semiotic phase. Mature language is obviously semiotic. Its ontogenesis, and perhaps its phylogenesis, may have involved stages in which vocalisations were first coordinations, alerts, adjustments, rhythms, summonses and responses. Representational signification would have emerged later. Several recent findings are consistent with it. In a comparison of interjections from 131 languages with nonlinguistic vocalisations, Ponsonnet et al. (2024) found that pain is the only affective experience among those tested whose vowel signature, open a-like vowels and wide falling diphthongs such as “ai”, is preserved from the nonlinguistic cry to the interjection across languages. The word for pain keeps the acoustic shape of the bodily event it accompanies, which is what one would expect if the vocal praxis was coupled to the algesic praxis before it became a lexical item. Dingemanse (2024) reviews evidence that interjections, long treated as marginal, are tools for conducting interaction and help children break into language. The developmental findings discussed in sections 3.4 and 3.6 point the same way: the infant’s expectation, at five months, that its vocalisations will alter what adults do (Elmlinger et al., 2023) and the relation between vocal turn-taking and neural synchrony at four to six months (Nguyen et al., 2023). A longitudinal study adds that caregivers’ contingent responses to babbling at five months predict vocal turn-taking at ten months, and that the growth of prelinguistic vocalising and turn-taking over this period predicts the child’s early language (Zhang et al., 2024). In these studies vocalisation coordinates conduct between two organisms well before it can be said to stand for anything. On the philosophical side, Bar-On (2021) argues that the search for precursors of linguistic communication needs an intermediate notion of pragmatics, broader than the Gricean one built on communicative intentions, so that pragmatic phenomena can be recognised where speaker meaning is absent. 3.11 Decoupling and the birth of the sign The sign is born of decoupling. When a vocalisation can occur without the original situation, a decisive transformation takes place. “Ai” can be said while remembering the pain, imitating someone, telling a story or warning of a danger. The vocalisation is then no longer only coupled to the present situation. It can evoke it. The path is coupling → stabilisation → decoupling → evocation, and it is here that the semiotic function properly begins. Deacon (1997) described a comparable passage when he derived symbolic reference from indexical relations that become progressively detached from the immediate context. In this theory a sign is a stabilised praxic form that has acquired the capacity to function decoupled from the original 12 Research, Society and Development, v. 15, n. 10, e38151051805, 2026 (CC BY 4.0) | ISSN 2525-3409 | DOI: http://dx.doi.org/10.33448/rsd-v15i10.51805 processuality and to reactivate, evoke or cognitively replace it. This is very different from presupposing the sign from the beginning. Signification, in turn, is not at first a mental thing attached to a word. It is a historical stage of the praxic relation, and can be defined as the capacity acquired by a communicative form to mobilise a processuality in the absence of its immediate occurrence. The semiotic function emerges from the history of coordination. Meaning does not at first cause efficacy; efficacy prepares meaning. In the traditional model the word works because it means. In the proposed model the vocalisation comes to mean because it has previously worked, repeatedly, within certain coordinations. The order is praxic efficacy → stabilisation → meaning, and not necessarily meaning → efficacy. This removes an infinite regress. The regress appears when it is said that to learn a word I need a concept, that to learn the concept I need another concept, and that to explain that one I need yet another. If the base is the coupling of praxis with praxis, the chain ends, or better, it begins before the concept. We do not need an object for everything, a representation for everything or a sign for everything. 3.12 Object, representation and characteristic as late products The last part of Hanna’s formulation speaks of “perceptual representations indicative of objects”. The expression presupposes objects, representations of objects and characteristics of those objects. These are exactly what should be explained. In the perspective proposed here the cognitive object is late, and experience does not begin with ready-made objects. Representation is late as well. Before there is a “representation of the lion” there is being chased, hearing, seeing, running, feeling fear and orienting oneself. The concept of representation may be imposing a photograph on what was originally a processual film. A large part of modern epistemology works with a similar scheme, the old theory of the ready-made object: external object → internal representation → concept. The present thesis proposes another: processuality of organism and world → lived experience and praxis → communicative reflexivity → objectification → form → representation and concept. The inversion is substantial, and it agrees with the critique of representation carried out by ecological and enactive approaches to perception (Clark, 1997; Gibson, 1979; Varela et al., 1991). Fuchs (2026) reaches a similar point from phenomenology and enactivism: putting experience into words partly undoes the unity of the lived situation, the separation of subject and object is secondary and becomes fixed through language, and a noun such as “table” names at first a practical dealing with the thing, from which the object is isolated only afterwards. Early representation creates artificial problems. Once the representation of objects is presupposed, questions follow: how do we classify these representations, how do we abstract what is common, how do we synthesise characteristics, how do we guarantee identity, how do we arrive at unity? Several of these problems may have been created by the initial supposition that delimited object representations were already there. The object concept can emerge from processualities, lived experiences, praxes, partial coagulations and communication. To abstract characteristics from objects therefore does not explain the first conceptualisation. It is an operation that becomes possible later. The characteristic is itself a late product. To speak of mane, tail, brown or four legs already presupposes that experience has been segmented. A characteristic is not necessarily a raw datum of perception. It is also a conquest of formalisation, obtained in the chewing of experience described in section 3.3, and so the Kantian account as presented by Hanna begins with yet another late product. Before the characteristic there is lived difference. Before “brown” exists as an explicit characteristic there is chromatic experience. Before “fast” there is the experience of displacement and temporality. Before “furry” there is a certain tactile and visual experience. Before “aggressive” there is a relational configuration that is lived. The order is lived experience → differentiation → stabilisation → characteristic, and not simply a characteristic that is already present followed by abstraction. A more suitable word than “generation” may be sedimentation. The concept is sedimentation and not fabrication. It 13 Research, Society and Development, v. 15, n. 10, e38151051805, 2026 (CC BY 4.0) | ISSN 2525-3409 | DOI: http://dx.doi.org/10.33448/rsd-v15i10.51805 sediments experiences, relations, actions, vocalisations, memories and communications, and afterwards remains available as a reusable form. This explains both its stability and its history. 3.13 The architecture in levels The complete theory can be set out in twelve levels, shown in Table 1. Table 1 - Levels of the genetic architecture, from processuality to propositional language. Level Designation Content 1 Processuality Organism and world in interaction. 2 3 Lived experience and praxis Intraindividual coupling Action, perception, affection. Motor, vocal, sensory and affective processes coupled within the organism. 4 5 Interindividual coupling Synchronisation One praxis modifies another. Recurrent configurations between subjects. 6 7 Verbal accompaniment Internalisation Vocalisations come to accompany experiential regularities. The accompaniment by the other becomes self-accompaniment. 8 9 10 Communicational reflection Objectification Formalisation The subject organises its own lived experience for the interlocutor. A relatively stable pole is extracted from the flow. Object and concept become reproducible. 11 12 Semiotisation Propositional language Decoupled forms come to represent and evoke. Forms can be articulated explicitly. Source: Authors. Table 1 shows that what the Kantian formulation takes as origin occupies the last four levels. Objectification, formalisation, semiotisation and propositional language are reached only after eight levels in which nothing needs to be represented. Levels 3 to 5 describe couplings and synchronisations that are documented in the first months of life (Nguyen et al., 2023; Trevarthen, 1979). Levels 6 and 7 correspond to the mother’s accompaniment and its internalisation, described in section 3.6 (Fernyhough & Borghi, 2023; West et al., 2022). Level 8 introduces the communicative volition and the chewing of experience for an interlocutor discussed in sections 3.3 to 3.5. The secondary self of section 3.7 is constituted at level 7, and levels 8 and 9 are its work. The problem is an inversion of priorities, which becomes visible when the two sequences are placed side by side (Table 2). Table 2 - Explanatory order in the Kantian formulation reconstructed by Hanna and in the processual theory. Step Kantian formulation Processual theory 1 2 Mental operation Rule World and process Organism 3 4 Synthesis Understanding Lived experience and praxis Coupling 5 6 Characteristics Representation of the object Synchronisation Communication 7 8 Concept Self-reflection Objectification 9 10 11 Form Signification Concept 12 Rule and reflective understanding Source: Authors. 14 Research, Society and Development, v. 15, n. 10, e38151051805, 2026 (CC BY 4.0) | ISSN 2525-3409 | DOI: http://dx.doi.org/10.33448/rsd-v15i10.51805 The two columns of Table 2 differ in causal and genetic order, and not merely in terminology. In the first column the concept is the seventh and last term, and everything before it is already mental and formal. In the second the concept is the eleventh term, and rule and understanding, which open the first column, close the second. 3.14 Consequences for epistemology These results modify epistemology itself. Knowledge ceases to be only a relation of subject to object. The basic structure becomes triangular: world, subject and other subject. Communication is not added to knowledge afterwards. It takes part in its formalisation (Tomasello, 1999). Objectivity and communication come close to each other again. Communication is a matrix of objectivity: when I have to make my lived experience communicable, I must take from it something stable, re-identifiable and shareable, which is what stays constant from one listener to the next (section 3.4), and this favours the formation of form. Form thus has a genealogy and is a communicational product. It does not fall ready-made from the understanding; we produce it because we need to recognise, repeat, communicate and compare. Form is processuality stabilised for cognitive circulation. A philosophy may reply that this is psychology and not logic. That separation is exactly what is in question. Formal validity can remain independent of psychological facts. The genesis of the forms that logic uses is nonetheless a real cognitive question, and to refuse the investigation for fear of psychologism does not solve the problem. It only excludes it by method, and a genetic epistemology undertakes to lift that exclusion (Piaget, 1970). To naturalise is not to reduce. No one is claiming that concepts are neurons or that truth is a brain state. The claim is that concepts have a history of constitution that takes place in real organisms. Studying that history leaves logical validity intact and explains how an organism becomes capable of logic. Cognitive Psychiatry occupies this interval. Between neurophysiology and formal epistemology there is a layer made of lived cognition, praxis, conceptualisation, self, inner language and communication. The layer has often been divided among psychology, linguistics, philosophy and neuroscience. Cognitive Psychiatry proposes to treat it as one continuous process. 3.15 The theses of the refoundation The argument can be condensed into ten theses, on the concept, language, the sign, the understanding, the rule, synthesis, representation, objectification, form and epistemology. They are stated in Table 3. Table 3 - Theses of the cognitive-praxic refoundation of epistemology. Topic Thesis Concept The concept is not originally the product of a specific mental operation of rule-governed synthesis. It is a late formal sedimentation of lived, praxic and communicative processualities, formed on neurophysiological and interindividual couplings that precede fully constituted representation and signification. Language Language does not necessarily arise when a vocalisation represents an object. It can begin in praxic couplings in which a vocal action modifies the behavioural organisation of another subject without the mediation of an explicit semiotic function. Semiotisation is later than communicative efficacy. Sign The sign is a late acquisition of a communicative praxis that has stabilised enough to become decoupled from the original situation and to evoke or replace it. Understanding The understanding should not be postulated as an original faculty that produces concepts. It can itself be understood as a higher level of stabilisation of praxic, communicative, reflective and conceptual capacities constituted earlier. Rule The explicit cognitive rule is later than processual regularity. The organism does not begin by applying rules to the world; through its praxis it incorporates regularities of a world in which it was itself constituted. 15 Research, Society and Development, v. 15, n. 10, e38151051805, 2026 (CC BY 4.0) | ISSN 2525-3409 | DOI: http://dx.doi.org/10.33448/rsd-v15i10.51805 Topic Synthesis Thesis Synthesis should not be presupposed as an original operation, because it already presupposes a formally distinguished manifold. Before elements to be synthesised there is processuality not yet cut up. Representation Representation is not necessarily the first mediation between organism and world. Before representation there is causal and praxic participation in the event itself. Objectification The cognitive object results from a late objectification by which a processuality is delimited, reidentified and made communicable. The external object can exist before it exists as a formal object within the mind. Form Form does not found process. Process progressively founds form. Epistemology An epistemology that begins with representation, concept, rule or proposition begins when a large part of the cognitive phenomenon has already taken place. A genetic epistemology must go back to the pre-objectal, pre-formal and pre-semiotic processualities from which these structures emerge. Source: Authors. Each thesis in Table 3 moves one category of the traditional architecture from the position of origin to the position of result. The thesis on form is perhaps the most economical statement of the inversion proposed. 3.16 Convergence with recent literature and limits The recent studies cited throughout this section were not designed to test the theory, and none of them, taken alone, establishes it. Their value lies in convergence. Work from different fields and with different methods places bodily coupling before representation (Ciaunica et al., 2021; Di Paolo, 2021), practical causal competence before objective causal knowledge (Goddu & Gopnik, 2024), vocal coordination before words (Elmlinger et al., 2023; Laudańska et al., 2025; Nguyen et al., 2023; Zhang et al., 2024), the mother’s verbal accompaniment of action before the child’s own narration (Elmlinger et al., 2025; Fernyhough & Borghi, 2023; West et al., 2022), the bodily cry before the interjection (Dingemanse, 2024; Ponsonnet et al., 2024), social interaction within concept formation (Borghi, 2022; Borghi et al., 2025; Fuchs, 2026) and a second-order level of self-regulation that appears late and in shared activity (Tomasello, 2024). The ordering is the one proposed here. Two limits should be stated. The pre-semiotic phase of language is a theoretical hypothesis: the available data show that vocal coordination is early and effective, and they do not by themselves show that it is free of any proto-meaning. And the phylogenetic extension of the sequence is offered as a possibility, since the evidence discussed is ontogenetic. Both points indicate where longitudinal and comparative studies could confirm or correct the model. 4. Final Considerations Hanna’s sentence concentrates the traditional architecture: the concept as mental product, generation, rule, synthesis, understanding, characteristics, representations and objects. The criticism developed here is that each of these categories has an earlier history. Before the mental operation there is lived experience. Before deliberate generation there is neurophysiological automatism. Before the rule there is processual regularity. Before synthesis there is a flow not yet cut up. Before the understanding there are praxis and interaction. Before the characteristic there is lived difference. Before representation there is causal participation. Before the internal object there is external processuality. Before the sign there is functional vocalisation. Before signification there is coupling. Before symbolic language there is synchronisation. And before the subject is able to narrate itself there is another person who accompanied its experience and who was progressively internalised. This reconstruction changes epistemology deeply. The concept ceases to be the starting point and becomes a result. The understanding ceases to be first cause and becomes late organisation. The reflective subject ceases to be an original endowment and becomes the internalised place of an interlocutor. The rule no longer imposes form on experience and becomes the explicit statement of regularities that were lived beforehand. Language no longer arises from the relation of sign to meaning and acquires 16 Research, Society and Development, v. 15, n. 10, e38151051805, 2026 (CC BY 4.0) | ISSN 2525-3409 | DOI: http://dx.doi.org/10.33448/rsd-v15i10.51805 a non-semiotic prehistory of praxic coordinations. Representation no longer mediates mind and world from the origin and becomes a later formalisation of a causal and lived relation that came first. The inversion can be condensed in one sentence: we do not begin by representing the world; we begin by taking part in it. Afterwards we synchronise, communicate, reflect, objectify, formalise, represent and conceptualise. Traditional philosophy usually begins its analysis with the last terms of this sequence. The proposal of Cognitive Psychiatry is to reconstruct the natural history that made those last terms possible. What is at stake is therefore more than the correction of one particular Kantian theory of the concept. It is a broader change, from the epistemology of ready-made form to an epistemology of the genesis of form. Future studies may test the sequence through longitudinal observation of early vocal and motor coupling and examine its phylogenetic counterpart. References Bar-On, D. (2021). How to do things with nonwords: Pragmatics, biosemantics, and origins of language in animal communication. Biology & Philosophy, 36(6), 1–25. Borghi, A. M. (2022). Concepts for which we need others more: The case of abstract concepts. Current Directions in Psychological Science, 31(3), 238–246. https://doi.org/10.1177/09637214221079625 Borghi, A. M., Mazzuca, C., & Tummolini, L. (2025). The role of social interaction in the formation and use of abstract concepts. Nature Reviews Psychology, 4(7), 470–483. https://doi.org/10.1038/s44159-025-00451-z Bruner, J. (1983). Child’s talk: Learning to use language. Oxford University Press. Ciaunica, A., Constant, A., Preissl, H., & Fotopoulou, K. (2021). The first prior: From co-embodiment to co-homeostasis in early life. Consciousness and Cognition, 91, Article 103117. https://doi.org/10.1016/j.concog.2021.103117 Clark, A. (1997). Being there: Putting brain, body, and world together again. MIT Press. Deacon, T. W. (1997). The symbolic species: The co-evolution of language and the brain. W. W. Norton. Di Paolo, E. A. (2021). Enactive becoming. Phenomenology and the Cognitive Sciences, 20(5), 783–809. https://doi.org/10.1007/s11097-019-09654-1 Dingemanse, M. (2024). Interjections at the heart of language. Annual Review of Linguistics, 10, 257–277. https://doi.org/10.1146/annurev-linguistics-031422124743 Elmlinger, S. L., Levy, J. A., & Goldstein, M. H. (2025). Immature vocalizations elicit simplified adult speech across multiple languages. Current Biology, 35(4), 871–881. https://doi.org/10.1016/j.cub.2024.12.052 Elmlinger, S. L., Schwade, J. A., Vollmer, L., & Goldstein, M. H. (2023). Learning how to learn from social feedback: The origins of early vocal development. Developmental Science, 26(2), e13296. https://doi.org/10.1111/desc.13296 Erlewein, D. (2026). Kant’s solution to the trilemma of concept formation. European Journal of Philosophy, Article e70060. https://doi.org/10.1111/ejop.70060 Fernyhough, C., & Borghi, A. M. (2023). Inner speech as language process and cognitive tool. Trends in Cognitive Sciences, 27(12), 1180–1193. https://doi.org/10.1016/j.tics.2023.08.014 Fuchs, T. (2026). The experiential basis of concepts: Integrating embodied and enactive accounts. Frontiers in Psychology, 16, Article 1710102. https://doi.org/10.3389/fpsyg.2025.1710102 Gibson, J. J. (1979). The ecological approach to visual perception. Houghton Mifflin. Goddu, M. K., & Gopnik, A. (2024). The development of human causal learning and reasoning. Nature Reviews Psychology, 3(5), 319–339. https://doi.org/10.1038/s44159-024-00300-5 Hanna, R. (2004). Kant e os fundamentos da filosofia analítica (L. S. Mendes, Trans.). Editora Unisinos. (Original work published 2001 as Kant and the foundations of analytic philosophy, Oxford University Press) Kant, I. (1992). The Jäsche logic. In J. M. Young (Ed. & Trans.), Lectures on logic (pp. 521–640). Cambridge University Press. (Original work published 1800) Kant, I. (1998). Critique of pure reason (P. Guyer & A. W. Wood, Eds. & Trans.). Cambridge University Press. (Original work published 1781/1787) Laudańska, Z., Babis, K., Kozioł, A., Szmytke, M., Marschik, P. B., Zhang, D., Malinowska-Korczak, A., López Pérez, D., & Tomalski, P. (2025). Context shapes (proto)conversations in the first year of life. Developmental Science, 28(3), Article e70018. https://doi.org/10.1111/desc.70018 Merleau-Ponty, M. (1945). Phénoménologie de la perception. Gallimard. 17 Research, Society and Development, v. 15, n. 10, e38151051805, 2026 (CC BY 4.0) | ISSN 2525-3409 | DOI: http://dx.doi.org/10.33448/rsd-v15i10.51805 Nedergaard, J. S. K., & Lupyan, G. (2024). Not everybody has an inner voice: Behavioral consequences of anendophasia. Psychological Science, 35(7), 780– 797. https://doi.org/10.1177/09567976241243004 Nguyen, T., Zimmer, L., & Hoehl, S. (2023). Your turn, my turn. Neural synchrony in mother–infant proto-conversation. Philosophical Transactions of the Royal Society B: Biological Sciences, 378(1875), Article 20210488. https://doi.org/10.1098/rstb.2021.0488 Pereira, A. S., Shitsuka, D. M., Parreira, F. J., & Shitsuka, R. (2018). Metodologia da pesquisa científica [e-book]. Editora UAB/NTE/UFSM. Piaget, J. (1936). La naissance de l’intelligence chez l’enfant. Delachaux et Niestlé. Piaget, J. (1970). L’épistémologie génétique. Presses Universitaires de France. Ponsonnet, M., Coupé, C., Pellegrino, F., Garcia Arasco, A., & Pisanski, K. (2024). Vowel signatures in emotional interjections and nonlinguistic vocalizations expressing pain, disgust, and joy across languages. The Journal of the Acoustical Society of America, 156(5), 3118–3139. https://doi.org/10.1121/10.0032454 Tomasello, M. (1999). The cultural origins of human cognition. Harvard University Press. Tomasello, M. (2024). An agency-based model of executive and metacognitive regulation. Frontiers in Developmental Psychology, 2, Article 1367381. https://doi.org/10.3389/fdpys.2024.1367381 Trevarthen, C. (1979). Communication and cooperation in early infancy: A description of primary intersubjectivity. In M. Bullowa (Ed.), Before speech: The beginning of interpersonal communication (pp. 321–347). Cambridge University Press. Varela, F. J., Thompson, E., & Rosch, E. (1991). The embodied mind: Cognitive science and human experience. MIT Press. Vygotsky, L. S. (1986). Thought and language (A. Kozulin, Ed.). MIT Press. (Original work published 1934) Wang, W. (2021). Kant on the formation of empirical concepts. Kant-Studien, 112(2), 195–216. https://doi.org/10.1515/kant-2021-0017 West, K. L., Fletcher, K. K., Adolph, K. E., & Tamis-LeMonda, C. S. (2022). Mothers talk about infants’ actions: How verbs correspond to infants’ real-time behavior. Developmental Psychology, 58(3), 405–416. Wittgenstein, L. (1953). Philosophical investigations (G. E. M. Anscombe, Trans.). Blackwell. Zhang, V. H., Elmlinger, S. L., & Goldstein, M. H. (2024). Developmental cascades of vocal turn-taking connect prelinguistic vocalizing with early language. Infant Behavior and Development, 75, Article 101945. https://doi.org/10.1016/j.infbeh.2024.101945 18