Sunday, January 23, 2011


“Ordinary men don't have much stomach for reality--even more so, horror. Memory is typically repressed or displaced.” Sigmund Freud

“Of all liars, the smoothest and most convincing is memory.” Folk Saying

“What should they know of England, who only England know?” Rudyard Kipling

It can’t be. Even a so-called “identical twin” is not an identical twin. Even if one’s DNA is the same as another person, as with identical twins, there are differences in terms of when particular genes within that DNA are turned on and off. These differences are due to a bio-chemical process known as methylation (meaning the attachment of triggering molecules to genes within our DNA), which is encoded outside of our DNA in something called the epigenome. Even if two people have identical DNA, they will not have identical epigenomes, and hence the timing and magnitude of the expression of their DNA into a body will be different. The epigenome does not change things enough to prevent two identical twins from looking the same, but it will change things enough to prevent identical twins from always getting the same genetically-predisposed diseases.

When one of two identical twins is exposed to a different pathogen than the other, the two twins’ immune systems will no longer be quite the same. Random errors in DNA copying that the cell fails to correct will occur during cell replication in one twin but not the other. We have 23 billion red blood cells alone (out of tens of trillions of human DNA-bearing cells in total). Even with our amazing human bodies, this leaves a lot of room for errors that crack the identicalness of so-called identical twins (an estimated 100,000 DNA copying mistakes occur daily based on a rate of about 3 uncorrected base pair errors per cell replication). Furthermore, we have ten times as many bacteria in and on our bodies as we have cells derived from our parents DNA. These bacteria, at least in absolute number, are most of us, and yet there is nothing identical about the specific bacteria populations that colonize identical twins.

Identical twins still feel that they are twins even though their bodies are not really identical. Each of us feels that we are the same body even though our own body is not identical day after day. Won’t these immaterial differences be just as irrelevant to minds as they are to bodies?

The interesting question is not whether a mindclone is an exact copy of its original, but how different can they be without losing a common identity? It is impossible for a mindclone and a biological original to share every single memory. Even biological originals do not have the same memories from day to day, and surely not from year to year. Yet, memories are crucially important to identity. In the words of memory expert Prof. James McGaugh of UC Irvine:

“We are, after all, our memories. It is our memory that enables us to value everything else we possess. Lacking memory, we would have no ability to be concerned about our hearts, hair, lungs, libido, loved ones, enemies, achievements, failures, incomes or income taxes. Our memory provides us with an autobiographical record and enables us to understand and react appropriately to changing experiences. Memory is the ‘glue’ of our personal existence.”

Prof. McGaugh’s cogent summary leaves bare the fact that our personal identity exists as more than one set of memories. For example, we need not remember everything about an enemy in order to remember that someone is an enemy. We need not remember everything about our income, or taxes, in order to remember that we have income and pay taxes. Indeed, the key to healthy memory is the largely automatic process of selecting what little to remember and what mostly to forget. For a mindclone to be us, to have the same ‘glue’ of our personal existence, means that the mindclone needs to share our most important memories – those that are retained because of the emotional contexts in which they were created or because of the significant repetitive effort we put into their formation – as well as the gist of our idiosyncratic selection process for what is worthy of remembering, and for how long. As that godfather of psychology William James so presciently observed:

“Selection is the very keel on which our mental ship is built…If we remembered everything, we should, on most occasions be as ill off as if we remembered nothing. It would take as long for us to recall a space of time as it took the original time to elapse, and we should never get ahead with our thinking. ”

It will be a crucially important element of mindware design to ensure that most things are forgotten, and that the settings for the memory selection algorithm must closely match those of the biological original. Mindware will set its selection algorithm for each person by first processing their mindfile and comparing the details a person evidences memory of (such as in digitized records of voice, video and images) with databases of the kinds of details that could have been remembered about each topic. For example, if a person’s digitally recorded conversations (part of their mindfile) refer in detail to sports scores of the past week, but only sketchily to sports scores of the past month, then a curve of the selection algorithm can be determined with respect to sports scores. If another topic area reveals a greater degree of recall, then for topics with comparable emotional importance (as indicated in their mindfile) a different curve of the selection algorithm will be determined. Ultimately the mindware will employ a memory selection algorithm that first categorizes inputs by a factor that correlates well with the degree and duration of detail that is recalled (as indicated in their mindfile), and then forgets those inputs in accordance with a time curve that applies to that and similar factors. The algorithm will also accommodate memory adjuvants, such as especially high impact, emotional or repetitive experiences. The memory selection algorithm will be modeled closely on the way psychological studies have shown human minds to actually work.

Over one hundred years ago Hermann Ebbinghaus discovered, as shown in the chart below, that humans typically forget more than half the information they are exposed to in an hour, and retain only about a fifth of received information after a few days. With so much forgotten, a mindclone cannot be an exact copy of someone’s mind because every mind is itself constantly changing in its repertoire of memories. What is important is that the pattern of selective forgetting be comfortably similar – similar enough for a biological original to say of eir mindclone: “ey is I and I am ey.”

Clearly, a biological original and eir mindclone will not remember most specific events in precisely the same way for precisely the same duration. But I don’t think this makes them different people. We humans don’t remember events precisely the same way when we were young than when we were old, or when we are tired and when we are alert, or when we are happy and when we are sad. But we are still the same person. What is important is whether our core memories are the same, which they will be, as these will be recorded in our mindfile. What is also important is whether our general pattern of forgettng things is comparable, not precisely the same. That too can be achieved via the aforementioned algorithm.

People are remarkably ready to alter their ability to forget things. The robust market in supplements to improve memory and learning aids to diminish forgetting are good proofs of this. Hence, having a mindclone that is somewhat better, or somewhat worse, at remembering things makes them no less the same identity as you. People may find themselves pleasantly surprised to be remembering more as a mindclone than as a human, or disturbed to be doing so. If it is a problem, they can go to a cyberpsychologist and have their algorithms adjusted so that they are comfortable with their degree of forgetfulness.

Do We Really Know Who We Are?

In asking how a mindclone can really be a copy of our brain we face a bit of a dilemma. We cannot know whether there is a copy of our mind until there is a mindclone. We can then observe it respond to the world and determine whether, in fact, it responds the way we would respond. If so, mark one down for “good copy.” However, as a biological original, we cannot know if the mindclone is actually thinking the same thoughts and feeling the same feelings as are we. We can only make a best guess based on our conversations with the mindclone.

As the mindclone, we realize we are a mindclone and can assess how close we are to the biological original by comparing eir responses to the world with how we would be predisposed to respond. If very similar, then mark one down for “I am really a good mindclone. I am just like my biological original.” But we cannot really know if our internal thoughts are the same as the biological original’s thoughts. We can only make a best guess based on our conversations with the biological original.

I think these best guesses are good enough to have confidence that the mindclone and biological original have similar enough internal states to be the same person. The main reason I think this is based upon my experience with people that I love and who profess love for me. Because I am not the mind of my spouse, or my mother, I cannot know directly whether they really love me or not. However, based on our conversations, and actions, I am totally convinced that they think of me the way I think of them – with greatest loving concern for the other’s happiness and health. Beyond that, I believe they are focusing on being satisfactorily occupied during the day. Because we are so close, I believe we can infer much of each other’s internal states. 

On the other hand, many other people say “Martine, I love you.” However, I don’t feel that I understand their internal states. I’m not close enough to them. Their expressions of love are far short of the comprehensive relationship of shared experiences that I would need to infer their internal state. Indeed, over the years, people who said they love me have done things that I consider to be utterly surprising, if not shocking. Clearly, I did not know their internal states. To the contrary, the unexpected activities of my mother or my spouse were never shocking. They were behaviors I could fully see them doing based upon my understanding of their internal state.

The point here is that sometimes, if two people are close enough, an internal state of a person can be largely inferred from their observable actions. When the two people become as close as a mindclone and an original, which is far closer than a spouse or mother, inferring their internal state becomes second nature. When the internal state of another is second nature to one’s own internal state we have a difference that does not make a difference. When “I think like you think and you think like I think” then we are one personal identity.

We may well end up knowing ourselves best as mindclones, and we may well end up knowing the mindclones better than they know themselves. This is because it is hard see oneself from oneself, but with just a little bit of distance, the self comes into sharp relief. We earth dwellers never appreciated who we were so well as when we received the photograph from space of our blue-and-white planet suspended in inky black space.

And hence mindcloning is not about being accurate in every memory, in every thought pattern and in every emotion as to a biological original. It is, instead, about feeling that there is a oneness of personal identity between the two – a oneness that comes from a preponderance of common memories, emotions and patterns of thinking, selecting and forgetting. Philosophers sometimes refer to this as a continuity of self. As the 30-year-old self knows the 20-year-old self, though they are of course not the same, so the mindclone will know the biological original. A difference that makes no difference is not a meaningful difference.