Metaphors of Memory: A History of Ideas About the Mind Book

This is a wonderful book which stretched my bruised capacity for attention and understanding. I was ok in the kiddie pool of the first few chapters but once we got into holography, interacting light beams, and then neural networks afterwards, I was running to catch up.

The basic idea is that memory, and many mental processes, are hard to access and experiment on directly. Because of that, the use of metaphor to guide thought and research is important, more so than in processes which are more directly accessible to empirical analysis.

The first chapter is a discussion of what metaphor is, and of the theories and history of thinking about metaphor (metametaphorology?) and of the ways that metaphor guides, illuminates, describes, but also hides and sometimes distorts research into and understanding of the invisible processes it is intended to describe.

I really liked this chapter especially and if I was sharing any part of the book as a kinda movie trailer it would be this because it works almost as a standalone essay. It also goes, with depth and detail, into the varying ways our use of metaphor can help and hinder, and makes a cogent defence or apologia of the necessity of metaphor, and therefore the importance of us trying to understand how we use it and what it does. Which comes back again in the last chapter.

Draaisma has a calm, regular writing style which I enjoy. Which is handy as his paragraphs are fucking HUGE. My mind did skitter a little on this, through I think that's more because its been wrecked by my phone and the internet. To read this I lay down on my rug and read in silence (though its summer and there's birdsong outside a lot, birds are one of the memory metaphors in the book), measuring time with an hourglass, and did a chapter a day. One to two hours per chapter, which is a very slow reading time for me. I don't think I could have managed to concentrate fully in any other circumstances.

"Diogenes of Apollonia assumes that memories are contained in a particular section of the body's airways; he took as evidence the fact that people breathe a sigh of relief when what they have been trying to recall finally occurs to them."

That's from the second chapter 'memory as writing'. Two things, first is that is exactly the type of lo-fi near humorous or quasi factoid element that weaves throughout the book and lets its quite dense tone and informational structure breathe a little (IRONY), second, one of the few criticisms I had of the book is that it starts its history *at writing*. There is nothing in here on oral cultures, which is pretty typical for histories since there is shit all information on that and you need to to a lot of 'reconstruction', essentially intelligent bullshitting, to put anything down. Plus oral cultures tend to not be massively introspective or object-analytical in ways that make it easy to talk about metaphor. BUT - its would be so fucking interesting to hear if there were common elements or interesting stories or anything at all about how oral cultures talked about memory. Many of them were really good at poetics and had powerful and innovative imaginations.

Anyway, there is an end to my own megaparagraph.

Megaparagraph is a surprisingly pleasing word to write and speak.

Anyway.

Did you know that Robert Hooke had his own theory of memory?

"So that there is as it were a continued Chain of Ideas coyled up in the Repository of the Brain, the first end of which is the farthest removed from the Centre or Seat of the Soul where the Ideas are formed; and the other End is always at the Centre, being the last Idea formed, which is always the Moment present when considered: And therefore as there are a greater number of these ideas between the present Sensation of Thought in the Centre, and any other, the more is the Soul apprehensive of the Time interposed."

Lovely. Also, he did calculations to work out, literally, how many 'ideas' an average human would have over a lifetime, and how many they would need to store, the first mathematical analysis of the data-bearing load on the human mind (I think).

At every age and every advancement in physical knowledge, minds seize on whatever is subtle, sophisticated, novel, complex, whatever forms the bleeding edge of comprehension, and riddle their way through the conception of it like fungus, infesting it and breaking it down into metaphors of memory.

These metaphors are used for analysis, in particular they spread certain ideas in a way academic texts cannot. Many are much more useful for spreading an idea or concept than they are for explicating it once all the details are regarded. Holography in particular, Draaisma is not too impressed by for its explanatory power, but it sure as hell reached and influenced a lot of people.

This guy, Carus, from a freaky non-mechanistic interlude in memory metaphors, sounds interesting;

"Carus's writings on landscape are also about the soul, just as conversely, his lectures on psychology took inspiration from landscape painting. Nothing was further removed from Romantic views on nature, organism or soul than the image of clockwork, which could be disassembled or replaced part by part at will. The Romantic metaphors and analogies referred to natural processes, to what grew and slowed without links or interruptions. Carus wrote in his first letter on landscape that even if science has dissected a plant fibre by fibre, cell by cell, that knowledge will still not be sufficient to make a single leaf."

Carus had some freaky ideas about skulls, and I think a bunch of other questionable 19thC stuff. Don't worry, we get back to the machines soon enough.

Its in the 19thC that we get some of the first big exciting problems of memory and one of the ways in which it is very unlike a machine, and of course we get it by doing terrible things to animals, specifically slowly slicing up their brains in little pieces so we can see what happens to them.

They 'gracefully degrade'. Brains break slowly, unlike machines where one thing splinters or fails and then there is catastrophic loss, brains that suffer damage retain core functions, to lesser and lesser degrees, as more and more damage is done to them. This seems to be related to the 'flock of birds' quality that brains have, important information is both linked to specific areas and, also, distributed.

So you can do utterly terrible things to a dog or a rat and its will keep up its basic movement algorithms almost right to the end. The same is true for us, memory structures can be highly resistant to local damage, especially core functions like moving the body through space and finding things.

It's also in the 19thC that the rise of technology, from phonographs to photographs (Daguerrreotypes) creates a fresh bloom of mechanistic metaphors.

One of the qualities of or increasing understanding of memory, from Hooke on, is the gradual 'pressing in' and closing of mechanistic or material thinking, with more and more things being explained. So in the medieval period everything is basically just the soul and this is a meat puppet its briefly trapped inside. Then Descartes ruins that idea. From that point on, the 'soul' or the point of no analysis, the unobservable point, in some ways keeps retreating deeper and deeper, becoming more and more ethereal, but never quite disappearing. It is like a hand closing a fist on a slippery thing, squeezing but the thing keeps shifting and going away and the grip can never be absolutely firm.

So we kick into post WWI where people start developing telephone exchanges and the proto-mathematics of electronic computers.

This is interesting;

"As with Tolman, in Hull's case, too, metaphors functioned even outside direct research as an ordering scheme. Good simulations consisted of 'hierarchies of control' in order to direct sub-mechanisms. Scientific theories were constructed as deductive hierarchies of laws, a reflection of the hierarchical structure of reality. Even the social structure which Hull imposed on his own research group seemed to be inspired by the representation of a hierarchical machine. At the top was a select group of theoreticians ('postulate makers'). They passed their work on to a lower echelon 'deducers' or 'logic-grinders', mathematicians and logicians who had to derive as many hypotheses as possible from the theoretical postulates. The lowest level was that of the 'experimenters', mainly graduate students who subjected the hypotheses to empirical testing. If with Tolman a scientific researcher still had the dignity of a cartographer, with Hull he was no more than a cogwheel in the hectic machinery of scientific industry."

I do absolutely despise everything about that situation but it is certainly 40k as fuck. All praise the Omnisiah.

And then of course, finally the Babbage Engine levels up and simply wipes the floor with every other machine on every other level so far as metaphor productions goes. There is a lot of interesting stuff here about the deep intertwining of ideas, intuitions and metaphors about the mind and the development of computing. They cross-pollinate and reinforce each other continually, like priests of paired gods.

"... Within the metaphor for man-as-an-information-processing-system there emerged in the 1960s an exchange between the vocabularies for the human memory and that of computers. Terms like 'input', 'read-in', 'encoding', 'back-up memory', 'working memory', 'storage', 'address', 'matching', 'over-writing', 'search' retrieval', read-out' and 'output' acquired a place in a common vocabulary. In a psychological theories they referred to hypothetical processes in memory, and in AI theories to mechanisms and structures for information storage in computers. This shared psychological and technical meaning was acquired in an exchange which in general was balanced: over and against the technical influence of psychology on computer jargon. A term like 'retrieval' was taken by computer science from memory psychology. Terms like 'address' and 'encoding' found their way into psychology from AI. The same applies for the distinction of concepts which emerged in the context of AI - such as that between location-addressable and content-addressable - had not occurred previously in psychology. In some cases theoretical terms even went back and forth. For example, the term 'back-up memory' was derived from computers and was included in memory theories to indicate the temporary storage of information. Conversely the 'memory' in 'backup memory' derived from psychology, so that psychology had a term which it had lent out, returned to it."

You see what I meant about the long-ass paragraphs? But you feel like you learnt something don't you.

Draaisma breaks into one of his rare bursts of poetics in the computer age, describing, better than I could the differences highlighted by the advance of technology;

".. The memory of the computer is too good. It's infallibility is its principal short-coming. Human memory is an instrument which, if the need arises, lies and deceives. It distorts, sifts and deforms, takes better care of some things than others. unlike the computer memory it disobeys commands. it does not bother about instructions to keep one thing and throw something else away, it behaves like the disobedient dog that Cees Nooteboom called it. Whereas circuits in a classical computer are under a central operating system which gives its commands step by step, the human brain seems to be acted upon by scores of impulses at once. Odours, emotions, movements, sounds, perceptions: the memory as a vibrating network of synchronous associations rather than a linear tract of stimulus-storage-reproduction. The computer plays its melodies one key at a time, albeit incomprehensibly fast; the human memory strikes whole chords."

All this leads us into near-modern times where holography (which I still do not intuatively understand, please don't explain it in the comments), and then nerural nets close the distance further, though the slippery thing gets even more slippy. Also, according to Draaisma, holography is something of the flaky over-rewarded celebrity of memory metaphors, often quoted, rarely drops a good album. He thinks it just brings together a bunch of already-present ideas & isn't that testable iteself. This was written in 1891;

"Each element [of the nervous system] - speaking figuratively - may be considered as a minute area intersected by an indefinite number of curves of different directions and orders. Thus a molecular commotion in any such area may run into the system along any one of the innumerable curves. In every such small fragment "the whole curve slumbers".

Neural nets take us ever closer, clooooser, so that the metaphor and the model seem to merge, but through they can to a lot of brainy things, the problem of consciousness persists in the 'homonculus' the question in any mechanistic memory or brain metaphor of who or what is doing the regarding or operating them machine. Who looks at the hologram? Or, who or what is interacting with the stuff in the neural net, making it a 'memory'? And as ever, a more complete metaphor and a finer model only leaves us with those ethereal missing qualities, the things the networks cannot do.

"Mnemosyne has not given the same gifts to everyone. The memory that she gave to you and others is subject to external investigation. The memory that she gave me has a subjective, personal access, a secret door from inside. The price of that privileged access is a stubborn mystery: I do not know how my personal, introspective experience is linked to the observable processes in my brain. in my memory I cherish an intimate but unfathomable possession."

Finally Draaisma takes us back to the beginning;

"The history of memory is Heraclitic: no single metaphor or theory repeats itself in identical shape."

But they do repeat themselves, and many ideas and memories *of* memory are forgotten and returned to again and again throughout the history of our thought. Hence, Draaisma seems to say, this book, a memory of memory memories -MEMORY CUBED MOTHERFUCKER.