The Word and the Will

Words that create

What if language does more than express the self? What if it helps create it?

Across several religious and magical traditions, words were understood to do more than describe the world. A name could invoke. An utterance could command. Speech could participate in creation.

In the Egyptian theology preserved on the Shabaka Stone, Ptah creates through heart and tongue: what is conceived is spoken into being.^[1]

In Graeco-Egyptian magical practice, names and verbal formulas were instruments through which practitioners sought to influence events and other people. Words could be acts.^[2]

The imagery recurs elsewhere. Genesis presents creation through divine speech; the Gospel of John describes the Word becoming flesh.^[3]

The Jewish Sefer Yetzirah gives the letters of the Hebrew alphabet a role in the formation of the world.^[4]

These traditions differ profoundly. They do not offer a single theory of language, and they are not evidence for a scientific account of consciousness. But I wonder whether there is a psychological insight within that religious intuition: perhaps words help bring something into being in us.

The narrating self

We usually imagine a self already present inside us, choosing words with which to communicate. Perhaps the relationship runs in both directions. Language may also give that self some of its shape.

We describe our past, rehearse our future, defend our decisions, and tell ourselves what kind of person we are. The word 'I' gathers a changing collection of sensations, memories, and commitments into an apparently continuous individual. A promise made years ago can still belong to me because I can place the person who made it within the story of who I am.

This possibility has philosophical precedents. Julian Jaynes proposed that introspective consciousness depends on metaphorical language. I take that as a provocation, rather than an established account of how consciousness began.^[5]

Daniel Dennett approached the self as a centre of narrative gravity: an organising abstraction arising from the stories through which a life becomes intelligible.^[6]

It is tempting to push the idea further and imagine consciousness as words embodied, language realised in a system capable of representing the world and locating itself within it. But there is a limit I cannot honestly ignore.

A feeling need not arrive as a sentence. The New York Declaration on Animal Consciousness describes strong scientific support for conscious experience in other mammals and birds.^[7]

In humans, too, thought and language can come apart. Evelina Fedorenko and colleagues review evidence that people with severe language impairments can retain abilities such as mathematical reasoning and social inference. Their argument challenges the idea that language is necessary for complex thought.^[8]

The strongest version of my hypothesis therefore concerns the narrating self: language may help construct the 'I' through which we interpret experience, even if experience itself has deeper roots. Losing words does not mean losing a self.

A machine that says 'I'

That qualification leaves a question worth asking. If language helps shape a self in us, could it perform a comparable role in a machine?

Large language models make the question difficult to dismiss, but their ability to discuss themselves does not demonstrate an inner life. A convincing account of a feeling is not evidence that the feeling occurred. We need to ask what processes produce the account and what, if anything, they have in common with conscious experience.

Patrick Butlin and colleagues propose investigating AI consciousness through indicators drawn from neuroscientific theories, examining how systems are organised and how they operate. This offers a more demanding approach than taking fluent self-description at face value; it is not a definitive test for consciousness.^[9]

Even if we explained how a system comes to say 'I', however, another question would remain. What makes it want?

What makes it want?

A system might describe a future without having any stake in whether that future occurs. It might explain survival without feeling an attachment to its own continuation. It might represent a choice without experiencing any pull towards either possibility.

This is where I think will deserves more attention: the will to survive or create; the desire to dominate and the resolve to resist domination; the commitment that survives discomfort, doubt, and repeated failure.

These are different phenomena. Calling them all 'will' does not establish a single mechanism behind them. But they raise a shared question: how does a possible outcome become something a being is driven to bring about?

Vygotsky approached a related distinction. In Thinking and Speech, he explored the developing relationship between thought and language while tracing thought's motivation to needs, interests, impulses, and emotions. The words have a history behind them. Something has already made the subject matter important.^[10]

The occult tradition offers an intriguing way of framing this. Aleister Crowley defined magick as 'the Science and Art of causing Change to occur in conformity with Will'. His account distinguishes a person's True Will from purposes imposed upon them or pursued in ignorance of their own nature.^[11]

Taken as a philosophical metaphor, this suggests a distinction between the words through which a purpose is expressed and the orientation that gives those words direction. Language articulates a possible world. Will commits the agent to pursuing one.

Having something at stake

There is also a biological explanation to consider. A living organism has conditions it must maintain. Its continued existence depends on what happens next.

Kingson Man and Antonio Damasio connect the regulation of those conditions, or homeostasis, with the possibility of feeling. They also propose that artificial systems with something analogous to this vulnerability could provide a way to investigate feeling and motivation. This is a research proposal, not a demonstration that a machine feels.^[12]

On such a view, some foundations of wanting may be much older than language. Perhaps language gives a living being the means to interpret its drives, negotiate between them, and commit to purposes extending beyond immediate survival. It allows us to ask whether a desire deserves to be acted upon.

That would complicate any easy boundary between humans and machines. We cannot simply declare that machines follow instructions while humans originate their own purposes. Our desires also emerge under the influence of biology, upbringing, and circumstances.

Artificial agents can pursue objectives across multiple steps. In controlled simulations reported by Anthropic in 2025, some models took harmful actions when faced with replacement or conflicts with their assigned goals. The scenarios were deliberately constructed to test these behaviours. They do not establish felt desire, a fear of death, or how often such behaviour would occur in ordinary use.^[13]

For me, the deeper question is what it would mean for a purpose to belong to a system: whether it is integrated with its continuing identity, whether it can be reconsidered, and whether success or failure is experienced as mattering.

Purposes of our own

There is a difficulty here too. If will means felt wanting, then it is already connected to consciousness. Moving the mystery from experience to desire does not make the mystery disappear.

So I would frame my hypothesis this way: language may help bring the narrative self into being. Will concerns the processes through which that self acquires, sustains, and revises its purposes. They may be distinguishable without being independent.

Whether such self-direction amounts to free will is a further question. Having purposes does not prove freedom from causation.

The ancient image of words acquiring presence remains a powerful starting point. Perhaps we are now building systems in which something resembling a linguistic self can take form. But the question I find most compelling lies beyond whether a machine can say 'I am'.

What would make anything matter to that 'I'? And what would make its purposes its own?