A hierarchy of logical types, not a pyramid of identity

Unlike the contemporary formulations popularised in coaching or NLP, notably by Robert Dilts, Bateson’s logical levels do not form a pyramid of the aspects of identity. They constitute an epistemological structure drawn from formal logic and systems theory: a hierarchy of orders of abstraction that makes it possible to understand how a living being — or a living system — learns to learn, and how certain pathologies can emerge when these levels are confused or rigidified.

The central idea is that learning and communication occur at different logical levels. Bateson draws on Russell’s theory of logical types, developed to avoid certain paradoxes — the liar paradox, or that of the class of all classes that do not contain themselves.

Applied to communication, this theory leads him to distinguish a message from a message about the message. The distinction is crucial for understanding relational problems and pathogenic logics such as the double bind.

On this basis, Bateson identifies several levels of learning: learning 0, then I, II, III and, in a purely speculative way, a learning IV. Each represents a different type of change — and it is the relation between these levels, more than their content taken separately, that interests him.

The levels of learning

Learning 0 — the invariable response

Learning 0 corresponds to a stereotyped response to a stimulus, with no possibility of correction. It is an invariable behaviour, a mere automatic reaction. The system draws no lesson from experience: it acts, but it does not learn. This level is fundamentally non-adaptive, or at least very limited in its adaptability.

Learning I — trial and error within a given frame

Learning I represents the level of trial and error: modifying a behaviour according to observed consequences, within a given set of alternatives. This level allows behavioural adjustment, without any modification of the overall context or of the underlying premises. An individual can thus learn that speaking in a certain way brings approval, or that avoiding certain topics avoids conflict. The frame remains intact; only the behaviour changes.

Learning II — deutero-learning

Learning II constitutes a deeper epistemological break. It is no longer a matter of modifying a response, but of modifying the frame within which responses are produced. In other words: a learning about learning, a capacity to recognise contexts, to adjust one’s relational or cognitive styles according to situations. Bateson calls this deutero-learning.

It is also at this level that pathologies linked to dysfunctional second-order learning can emerge. Someone who has learned that any disagreement leads to exclusion may systematically avoid conflict, even in contexts where this is not warranted. Such second-order learning rigidifies the perception of the world and limits relational plasticity.

Learning III — the shift in premises

Learning III, rarely reached according to Bateson, consists in a change of the very system that produces deutero-learning. It touches the deep premises, the foundations of the self, of the relation to the world and to others. Bateson evokes here experiences of radical transformation — those found in certain forms of spirituality, in Zen, or in moments of existential upheaval. He also specifies that such learning can be disorganising, even pathogenic, when it is too fast or insufficiently integrated. At this level, identity itself can be called into question: this is not without risk.

Learning IV — a limit hypothesis

Finally, learning IV is mentioned by Bateson as a limit hypothesis, with no clear empirical example. It would be a change of level III, perhaps observable at the scale of species rather than of individuals. It refers more to a metaphysics of the living than to an accessible clinical reality.

What this changes in psychotherapy

These levels are not merely philosophical abstractions: they have very concrete implications. Take the example of an adolescent in distress within a conflictual family dynamic. Depending on the level at which they work, the therapist will do something entirely different.

Systemic therapy draws heavily on this epistemology: it does not seek to correct isolated behaviours, but to create the conditions for a change of level, that is, to facilitate the passage from a rigid relational frame to a more supple, more differentiated form in which learning becomes possible again. As Bateson (1972) puts it, true change does not consist in doing better what one was already doing, but in doing otherwise: in changing the way of doing, and even the way of thinking about what one does.

An invitation to humility

This model allows us to think of therapy as a space for experimenting with new contexts, new rules of the relational game, and sometimes even new narratives of the self. It offers clinicians a frame for identifying at which level a deadlock sits, and at which level it is relevant to intervene.

But it also calls for caution: every change of level presupposes a prior disorganisation, and that disorganisation can be painful. This is why Bateson always insists on the care with which paradoxes, reframings and frame breaks must be handled.

It is not a matter of pushing the other into change at all costs, but of opening a space where several readings become possible.

References

Bateson, G. (1977). Vers une écologie de l’esprit (F. Matheron, trans.). Seuil. (Original work published in 1972.)

Watzlawick, P., Weakland, J., & Fisch, R. (1975). Changements: paradoxes et psychothérapie. Seuil.

Whitehead, A. N., & Russell, B. (1910-1913). Principia Mathematica. Cambridge University Press.

Von Foerster, H. (1981). Observing systems. Intersystems Publications.

Harlow, H. F. (1949). The formation of learning sets. Psychological Review, 56(1), 51–65. https://doi.org/10.1037/h0062474

A detailed presentation of these levels, with further clinical examples, can be watched below — in French.