The transformative power of speech

In French culture, the written word prevails over spoken speech, which is held to be fleeting and less solid. From years of facilitating groups, I have come to see that this hierarchy is false. Living speech is a mode of thought in its own right, one that sometimes transforms more deeply than writing. But it has to be organised, because genuinely shared speech is never a given: it calls for work on the setting, on permission and on recognition, and that is what I want to examine here.

A persistent belief

In the French intellectual culture I have inherited, the written word prevails over speech. A written text is serious; a spoken discussion is fleeting. A book carries authority; a conversation passes. A thought shaped by writing has value; a thought feeling its way as it is spoken is held to be less solid. This hierarchy organises our institutions of knowledge: one does not become a philosopher through speech, one becomes one through published works. One does not earn a doctorate through conversations, one earns it through the deposited manuscript.

For a long time I shared this belief without ever putting it into words. And then, from facilitating groups, collective intelligence workshops and sessions of thinking together, I realised it was false. Spoken speech is not a degraded version of writing. It is a mode of thought in its own right, one that can be more transformative than writing, provided it is taken seriously. And our era, which is seeing the oral return on a massive scale through podcasts, video calls, voice messages and conversations with voice-based artificial intelligences, compels us to rethink this hierarchy.

Performative speech

In a series of lectures given at Harvard in 1955 and published posthumously as How to Do Things with Words (1962), John Austin made a major conceptual break. He showed that not all speech is descriptive. Some utterances do not describe the world; they modify it. When a mayor pronounces “I now declare you joined in marriage”, they are not describing a marriage that has already taken place: they bring the marriage into being through their speech itself. When a judge says “The session is open”, the session begins with those words, and not before. When someone says “I promise”, the promise exists from the moment it is uttered.

Austin calls these utterances performatives and distinguishes them from constatives, which describe a state of the world. He shows that performatives are neither true nor false in the usual sense: they are successful or unsuccessful, depending on whether the conditions for their efficacy are met. These conditions are not only linguistic: they include the authority of the speaker, the institutional context and the real commitment of the one who speaks.

This analysis has two major consequences. The first is that speech can be an act that modifies the world, not merely the trace of an act that took place elsewhere. The second is that analytic philosophy, which until then had centred on the truth value of utterances, had to acknowledge that a considerable part of our linguistic life escapes that criterion.

In Problems in General Linguistics (1966), Émile Benveniste extends this analysis by showing that subjectivity itself exists only in language. “It is in and through language that man constitutes himself as a subject”, he writes in the chapter entitled Subjectivity in Language. The I exists only because there is a language that allows someone to designate themselves by this pronoun, and which at the same time makes possible the you they address. Outside language there is no I and no you: there are biological organisms. Speech is what institutes subjects as subjects.

Vygotsky and inner speech

Lev Vygotsky, a Soviet psychologist born in 1896 in the Russian Empire, in Orsha in Belarus, and who died in Moscow in 1934, made a decisive contribution in Thought and Language (1934) from the field of developmental psychology. He shows that thought and language are not, in the child, two separate processes that meet at a certain moment. They develop in a common movement, in which each stage of language enables new forms of thought, and each cognitive acquisition transforms the use of language.

Vygotsky shows that inner speech, that is, the silent thinking we take to be the normal mode of thought, derives from social speech. The child first thinks by speaking aloud, then gradually internalises this speech until able to carry it on silently. Adult thought keeps the trace of this social origin: to think is essentially to speak to oneself, using the structures learned in speech with others.

This analysis has a counter-intuitive effect. If inner thought derives from social speech, then shared speech is not a degraded version of pure thought; it is, on the contrary, the original matrix from which individual thought is derived. What we usually call reflecting is internalised speech. What we call speaking together is, in this perspective, a more fundamental activity, one that reactivates the originary conditions of thought itself.

Gesture and speech

In Gesture and Speech (1964-1965), André Leroi-Gourhan situates this power of speech within a longer history, that of hominisation itself. He shows that language and the tool developed together, carried by the same freeing of the hand and the face over the course of evolution. Speech is not a faculty that was added to an already constituted body: it took shape in the same movement as upright posture, the making of tools and the social life of the first human groups. To think of speech as an act of the body in society, and not as a mere vehicle for ideas, draws on this history. What living speech mobilises of the body, of gesture, of rhythm, is not an archaic remnant from which writing has freed us; it is the very ground in which language grew.

What is transformed in shared speech

These analyses converge to give a framework to an experience many people know: there are conversations that transform more deeply than years of reading. An intense discussion, over a few minutes, a few hours or a few days, can change what we think, what we are, what we become. This transformation is not only intellectual: it modifies our relation to the world, to ourselves and to others.

Several features of shared speech account for this power.

• It unfolds in the presence of a witness. When I speak to someone who listens to me, I say things I would not say to myself alone, because the presence of listening calls forth speech. This speech, in turn, lets me discover what I think. The subject of the speech and the thinker of the thought are, in conversation, coming into being together.
• It is interactive. My interlocutor reacts, asks, contests, extends, which forces me to be more precise, to reformulate, to recognise what I had not seen. I do not think alone, I think with the other, in a movement where thought is co-constructed. In his analysis of dialogism, Mikhail Bakhtin showed that all authentic speech is dialogical: it is addressed to someone, it anticipates their replies, it takes up their formulations.
• It is alive. Rhythm, tone, silences, hesitations, laughter, changes of voice all carry a meaning that writing cannot carry. The same sentence, said with a certain intensity or with a certain lightness, does not do the same thing. Living speech mobilises bodily and affective dimensions that writing loses.
• It is situated. It unfolds in a place, at a moment, among people who have a history, and this situation gives it its singular relevance. The same thing said to other people, at another moment, would not do the same thing, because it would not be set within the same web of relations.

Collective intelligence as a device of speech

When I lead collective intelligence sessions, what I mean by that is an organisation of shared speech so that it becomes transformative. This calls for several things.

• A clear setting. A setting is not an arbitrary constraint, it is a structure that makes speech possible. One needs to know who speaks, to whom, about what, within what time, and by what rules. This setting, laid out explicitly at the start, then frees speech. Without a setting, speech scatters, is monopolised, or falls silent.
• Legitimation through writing. When I take notes, in the form of a mind map or a concept map, of what the participants say, writing does not replace speech, it legitimates it. Seeing one’s speech taken up, transcribed, made visible to others, transforms the relation each person has to their own speech. What I said counted, since it is written down. This validation through writing turns the speech of an individual into speech taken seriously by the collective.
• The circulation of speech. No one must monopolise, but no one must be compelled to speak either. Circulation happens through invitation, through solicitation, through respected silence. In a well-facilitated group, people who spoke little in the collective at the start begin, little by little, to dare more. This is not magic, it is the effect of a device that makes speaking up less risky.
• The move to individual writing. After the collective speech, which has set the common framework, each person can write for themselves. But this writing is no longer what it would have been without the prior speech. It is nourished by the words heard, authorised by the words said, structured by the framework built together. Collective speech and individual writing are not opposed, they nourish one another, in a cycle where each stage gains from existing thanks to the previous one.

These devices are not facilitation techniques among others. They mobilise what Vygotsky, Benveniste, Bakhtin and Austin theorised: speech as an act that institutes subjects, transforms thoughts and brings collectives into being.

The concept of instituting speech

I therefore propose to coin the concept of instituting speech to designate this dimension of speech that does not describe a prior state of the world but brings into being something that did not exist before it.

This concept takes up elements of Austin’s performative, of Benveniste’s subject, of Émile Durkheim’s collective effervescence and of Bakhtin’s dialogism. But it articulates them on practical ground: that of what happens in the living conversations that transform, in workshops of thinking together, in talking therapies, in citizen deliberations, in the love encounters that change a life.

Four features distinguish instituting speech from merely informative speech.

• It brings into being. It does not only say what is, it institutes what was not. The subject who speaks is constituted in their speech, the bond established between the speakers is built as it is spoken, the common thought emerges as it is said.
• It is situated. It could not take place elsewhere, at another moment, among other people. Its relevance depends on the singular situation in which it unfolds.
• It is reversible in its coming about. It can fail, not take, be missed, and is never guaranteed. This precariousness is its very condition: it is because it could fail that its success is an event.
• It is irreducible to writing. One can transcribe what was said, but the transcription does not contain what happened. The best transcriptions of therapeutic sessions or of transformative conversations always leave a remainder that cannot be put into written words. This remainder is not a technical flaw of transcription, it is constitutive of what living speech is.

The difficulty of making speech democratic

It would be an illusion to conclude from the above that it would be enough to start speaking together again for speech to become instituting. To start speaking again, yes, but to speak in a democratic way, which is quite another matter. Genuinely shared speech is not decreed by gathering people in a room and passing the turn to speak around. It calls for precise methodological work, demanding, never acquired once and for all, and this is probably the most important point when one wants to move from analysis to practice.

The first obstacle is that taking turns is not enough. One can give the floor to each person, in turn, and obtain only a series of formal contributions in which people say the bare minimum. Distributing speaking time fairly is necessary, but it is the easiest condition to meet, and the least decisive. What is at stake before the distribution of speech is the very possibility, for a person, of feeling entitled to speak.

This entitlement does not depend on the facilitator alone. In Language and Symbolic Power (Ce que parler veut dire, 1982), Pierre Bourdieu showed that all speaking unfolds on a linguistic market where not all speech is worth the same, and where each speaker anticipates, most often without knowing it, the value their speech will have in the eyes of others. The stage fright, the shyness, the silence of those who would in fact have something to say are not personal failings: they are the effects of this market, which discourages in advance those who feel distant from legitimate language. A person who has always been told, through a thousand channels, that their way of speaking was not the right one does not censor themselves out of shyness in the psychological sense: they have internalised that their speech is worth less, and they act accordingly. No fair distribution of speaking turns corrects this on its own.

One must therefore work upstream on the conditions that make speech possible. Several methodological gestures fit together.

• See to it that each person has been able to speak, and keep enough time for it. The facilitator’s role includes watching who has not yet spoken, and keeping enough time to invite, without forcing, those who have not dared. I regularly observe that these people then take the floor, and that they have important things to say. Had they not been invited, they would not have come forward, and the collective would have lost what they brought. The invitation is not a courtesy, it is a condition for speech to truly circulate.
• Build a setting that authorises rather than a setting that monitors. A reliable setting signals to each person that they have the right to come forward, and that what they say will be received without being sanctioned by rejection. To express oneself, especially when one does not know in advance what will come out of oneself, is to expose oneself to the group’s judgement, and the group’s judgement touches on something vital, because social exclusion is experienced by the organism as a threat. It is this guarantee of welcome, more than any encouraging words, that makes speaking up possible.
• Respect people’s dignity and cultural rights. If someone has always felt that what they liked, what they knew, what they thought was not legitimate, they will not grant themselves the value needed to speak. They will not even be in contact with the value of their own speech. Recognition by others is what makes possible the recognition of oneself by oneself, and it is this self-recognition that makes possible engagement in what is common. To put cultural rights into practice, here, is to recognise what each person brings with them, so that they allow themselves to set it down.
• Hold the facilitator themselves in a non-overarching position. A device that advocates cooperation but that the facilitator conducts in a top-down or condescending way infantilises those who listen, and contradicts in its form what it states in its content. Before my public talks, I often make clear that the role I occupy, that of carrying the public word, could be occupied by any person present, who would enrich the others with their knowledge. Letting go of power is work always to be taken up again, because power intoxicates and cuts us off from the other.

This work is never achieved once and for all. It is redone with each group, in each session, with the people who are there. Democratic speech is less a state than a path, a conquest that calls for attention to context, for time, and for the acceptance, on the part of the one who facilitates, of being displaced by what arises. It is on this condition that shared speech becomes instituting, and not by the mere fact that one has started speaking again.

Contemporary scientific implications

Contemporary neuroscience provides empirical elements that meet this philosophical analysis. Brain imaging studies show that speaking together activates specific cerebral areas, distinct from those activated in solitary reading or writing. Research on mirror neurons, since the work of Giacomo Rizzolatti in the 1990s, suggests that understanding another person passes through a bodily simulation of what that person feels and expresses, a simulation largely carried by speech in presence.

On the side of psychotherapy, the American psychologist Bruce Wampold, in The Great Psychotherapy Debate (2001, second edition with Zac Imel in 2015), synthesised a large number of meta-analyses to defend what he calls the contextual model. The finding that emerges is that the efficacy of a therapy depends less on the technique used than on the quality of the spoken relationship between therapist and patient, what research calls the therapeutic alliance. Shared speech, when it takes, heals; techniques without that speech heal little.

More recently, research on the cognitive and psychic effects of conversations at a distance, notably by video call, has brought to light what is lost when speech unfolds without physical co-presence. Without going so far as to say that all disembodied speech would be inferior, which is false for the reasons developed in the previous article on vocal presence, these studies confirm that speech in co-presence mobilises dimensions of the subject that are not mobilised otherwise.

What this analysis entails for our institutions, our schools and our democracies

This analysis has practical bearing for our era, in several domains.

• In our institutions, it calls for valuing spaces of living speech. Schools, universities, administrations and companies have massively shifted their exchanges towards digital writing: emails, platforms, shared documents. This migration has its reasons, but it has a cost, the loss of instituting speech. Organisations that no longer set aside time for speaking in common lose in collective intelligence, in cohesion, in capacity for invention. Reintroducing devices of living speech is not a step backwards, it is a condition for these organisations to keep thinking.
• In learning, it calls for reconsidering the place of speech. Classical pedagogy favours transmission through writing: lectures, textbooks, written exams. Active pedagogies, from John Dewey to Paulo Freire, insist on the contrary on shared speech as the matrix of learning, and contemporary research confirms this intuition. Teachers who reintroduce structured speaking time into their classes often obtain better results than those who stay with frontal transmission.
• In politics, it calls for thinking of speech as a condition of democracy. Contemporary democracies suffer from a crisis of deliberation: the spaces where citizens truly speak together have shrunk, in favour of one-way media discourse and often sterile online exchanges. Reconstituting spaces of instituting speech, in devices such as citizens’ assemblies, consensus conferences and popular assemblies, is not a luxury for anxious democrats, it is a condition for democracy to remain alive.
• Faced with artificial intelligences, it calls for renewed attention. These systems imitate speech without having its instituting dimension. An artificial intelligence does not constitute itself as a subject by speaking to me, and it does not bring a subject into being in me either, at least not in the same way as true human speech. This difference is not a devaluation of AI, which can render real services, but it matters for distinguishing types of speech and their effects.

Instituting speech is a fragile good. It is not preserved the way books are preserved, it exists only when it is practised. An era that no longer practises it loses it. And in losing it, it loses one of the dimensions through which subjects are constituted and thoughts are engendered. To recognise this dimension, to name it, to defend it, is one of the tasks that philosophy can take on today.