Memories Are Made of This : IN THE PALACES OF MEMORY: How We Build the Worlds Inside Our Heads, <i> By George Johnson (Random House: 22.95; 246 pp.)</i>

“History is not what you thought,” W. C. Sellar and R. J. Yeatman told us in “1066 and All That.” “ It is what you can remember. All other history defeats itself.”

“Is knowledge knowable?” the philosopher Woody Allen asks. “If not, how do we know this?”

The notion that our brains are physically changed--sometimes momentarily, sometimes permanently--by the formation of new circuits caused by the exposure to every book we read, conversation we have, look out the window we make, is so remarkable to George Johnson that “for the past three years it has been difficult for me to maintain much of an interest in anything else.” That may be a loss to his colleagues at The Week in Review section of the New York Times, where he is an editor, but it is a boon to anyone who has wondered, as he has, “How is it that memory leaves its mark so that we are able to carry the past inside our heads?”

“In the Palaces of Memory” is one of those rare books that explain to the layman in clear detail what is happening in a complex branch of science without trivializing the subject. Johnson also has the good sense to know that the scientists often are as interesting as the science, and his book is filled with wonderful characters whose personalities are revealed in their routes of inquiry and vice versa.

The title is drawn from the mnemonic device taught by 16th-Century Jesuit Matteo Ricci, who took to China a memory system used in the West since the days of Aristotle. To improve their power of retention, Ricci taught people to construct huge buildings in their minds, memory palaces, and to place a specific item in a particular room. After years of practice, the person could shut his eyes and mentally go to whichever room he needed for the information he knew was kept there.

Sentient beings are like Janus, the two-headed god. We look forward, absorbing the new, but we must look back to place it in perspective. Information without memory is noise.

As you read and ponder this, some of the tens of billions of information-processing cells called neurons that are part of the brain’s composition are being goosed by tiny electrical impulses. That juice squeezes through a treelike structure with thousands of branches called a dendrite, which transports all sensory information. Some of this news stimulates the neuron, some inhibits it.

If the neuron’s attention is more inhibited than stimulated, the good news will not be brought from Ghent to Aix. If the neuron’s attention is more stimulated than inhibited, then it fires a jolt of its own down a stalk called an axon. The axon in turn feeds through junctions called synapses into the dendrites of other cells, that jolt ferried across the synaptic gap by chemicals called neurotransmitters.

Rather like the delivery boy who left this newspaper on your doorstep this morning, the neurotransmitters drop the message on structures called receptors that are part of a dendrite. As you recognized the newspaper when you opened the door, so do the receptors recognize the transmitting molecules.

A single neuron can receive messages from thousands of other neurons; its axon can branch repeatedly, sending messages to thousands more as this process of chemical signals being converted to electrical signals, then back to chemical signals, proceeds at uncomprehensible speed through circuitry unimaginably complicated.

When the brain is exposed to a new event--these words, Michelle Pfeiffer’s face, a scent, a sound--a distinct pattern of neurons rev up and form a unique circuit, an engram, that acts as a symbol for this news from the outside world. When next these words, Michelle Pfeiffer’s face, that scent, that sound are encountered, the circuit is activated once again: A memory is recalled.

But how much of the initial one? Memories fade. How much do you remember of your last trip to the grocery store or the second hour of your wedding reception? Yet the essential part remains, a set piece on which to lever more recollection from the depths.

The ancient Greeks assumed the mind was centered in the heart, not the head. We know differently, at least insofar as the location of cells to poke around in that contribute to memory, but just what do we know? The irony of having a mind capable of learning, inquiry and almost infinite storage is that although we have turned that mind in on itself to divine its own workings, the mind has yet to fully reveal itself to itself.

Though not from lack of effort. Johnson, who also wrote “Machinery of the Mind: Inside the New Science of Artificial Intelligence,” here focuses on the search for the secrets of memory in something of a three-dimensional fashion by walking the reader through the minds of a trio of the leaders in that hunt, each of whom approaches from a different direction:

Neurobiologist Gary Lynch, “who with his gapped teeth, curly hair, and mischievous eyes . . . sometimes looks like a grown-up version of Alfred E. Newman, right down to his ‘What, me worry?’ smile”; physicist Leon Cooper (who shared the Nobel Prize in 1972 for his part in inventing the Bardeen Cooper Schrieffer (BCS) theory of superconductivity), “a theorist who spends a great deal of time theorizing about what it means to theorize”; and philosopher Patricia Churchland, “a materialist, who believes that everything is made of matter and energy, as well as a reductionist, who believes the mental states and brain states are one and the same . . . (and who) is resigned to the possibility that neuroscience may one day subsume philosophy.”

Lynch, a senior professor at the Center for the Neurobiology of Learning and Memory at the University of California’s Irvine campus, seeks to understand memory as we understand, say, digestion--as something biological--by studying neurons excised from the brains of rats. He floats the tissue in petri dishes and measures the tiny voltages the neurons emit. Then he slices it into cross sections less than a millionth of a meter thick, photographs them with an electron microscope, and looks for the trace that is left when an event is recorded inside us: meteorite-impact craters on a sub-cellular level--or, as Lynch puts it, “This enzyme eating that protein causes memory.”

Cooper, a professor at Brown University, uses computer simulations to try to explain how billions of neurons interact to generate the mind.

Churchland, who teaches at the University of California’s San Diego (actually La Jolla) campus and is the author of “Neurophilosophy,” has studied neurobiology and tries to appeal both to philosophers who know nothing about neuroscience and neuroscientists who know nothing about philosophy.

These and other people and their ideas make a fascinating and thought-provoking book. We see these inky marks on the page that we agree mean the sound of a certain letter and which form larger clumps we call words on whose meaning we have also agreed. But the clumps and the words are only metaphors-- dog is not a four-legged animal, only what we agree stands for the general conception of a hairy beast with bad breath (each specific connotation--terrier, spaniel, is another metaphor, and another abstraction).

The millions of chemical-electrical-chemical reactions are repeated (with variations) in our brains again and again, building new rooms to store the memories that allow us to think.

Maybe soon, we’ll know how. With luck, Mr. Johnson will tell us about it.