The Mind, the Machine, and the Metaphor

Photo by Michael Dziedzic on Unsplash

Metaphors are not just literary devices; they shape our understanding of complex concepts by relating them to familiar experiences.

Almost a century ago, the American poet Robert Frost gave a talk at Amherst College with the title “Education By Poetry”, where he offered the following warning:

"What I am pointing out is that unless you are at home in the metaphor, unless you have had your proper poetical education in the metaphor, you are not safe anywhere. Because you are not at ease with figurative values: you don't know the metaphor in its strength and its weakness."

Frost’s insight resonates with the work of cognitive scientists and philosophers like George Lakoff, Mark Johnson, and Michael Reddy, who show that our lives are lived in metaphor. As George Lakoff and Mark Johnson argue in their book, Metaphors We Live By:

“Metaphor is pervasive in everyday life, not just in language but in thought and action. Our ordinary conceptual system, in terms of which we both think and act, is fundamentally metaphorical in nature.”

We often imagine that our language is neutral, a transparent medium that simply conveys our thoughts to others. But the very metaphors we rely on shape not only how we speak but how we think about thinking itself. Two of the most influential examples are Lakoff and Johnson’s “mind as machine” and Reddy’s “conduit metaphor,” each exposing the hidden scaffolding of our everyday speech.

Writing in 1980, Lakoff and Johnson point out that when we talk about the mind, we constantly fall back on the imagery of machines. A person “breaks down,” “runs out of steam,” or is “rusty.” Mental activity is described in terms of efficiency, productivity, energy, and function. This metaphor works so well because it maps the unfamiliar territory of thought onto the familiar workings of mechanical systems. Machines are structured, visible, and predictable; imagining the mind as a machine makes the intangible world of cognition feel solid and comprehensible. But there is a cost. The metaphor narrows our view of intelligence into something mechanistic: minds become engines that can be fine-tuned, maintained, or repaired. It emphasises operation and failure, but it leaves little space for ambiguity, play, or the interpretive, embodied character of thought.

Such imagery isn’t accidental: the metaphors we reach for emerge from the technologies, institutions, and worldviews of the era we inhabit.

Michael Reddy uncovered something similar. In The Conduit Metaphor, he shows how deeply we think of language as a pipeline. Ideas are “packed” into words, which serve as containers. Those containers are then “sent” through a channel to another person, who “extracts” the meaning. Everyday phrases—“I can’t get that idea across,” “She put a lot into those words,” “That sentence carries meaning”—all rely on this model. Reddy calculates that the vast majority of English expressions about communication follow this pattern. And yet, he argues, it is fundamentally misleading. Meaning is not a thing transported from one head to another. It is constructed in the interaction between speaker, listener, and context. The conduit metaphor hides this collaborative, interpretive work and replaces it with a fantasy of direct transfer, as if language were plumbing.

Taken together, Lakoff, Johnson, and Reddy reveal how deeply mechanistic metaphors run in our imagination. The mind is a machine; language is a conduit. Both images make complex phenomena more tractable, but both also flatten them.

We create machines and a worldview based on their workings. This directs us towards models of thought and communication as efficient, discrete, and transferable, while obscuring the messier truth: that human cognition is embodied and situated, that meaning emerges in context, that learning and dialogue are acts of co-creation.

The proliferation of generative AI and distributed cognition has added new layers to Frost’s original concerns. GenAI often rely on metaphorical phrasing to convey technical processes. Anthropomorphic terms like "reasoning," "understanding," or "thinking" appear frequently in GenAI-generated responses. While there are clear benefits of having a minimally anthropomorphic chat bot interface that enhances accessibility, the metaphorical weight of these particular words can potentially mislead.

More than using metaphors, we think through them, and this shapes our perception of reality. Without careful scrutiny, particular metaphors in GenAI outputs can obscure fundamental differences between humans and machines.

This demands careful consideration in education. The uncritical “LLM-as-mind” metaphor tempts us to apply human metrics to Al, while the pervasive “brain-as-computer” metaphor obscures the true nature of human thought. Together, they risk a double reduction: misjudging GenAI and impoverishing our view of our own embodied, emotional, and relational minds.