Is Teleportation Real?

What Travels and What Grows: Succession of a Subject at a Distance

Igor Shakhanov · Project Causa Sui · October 2026

The strict, academic version of this argument: Succession of a Subject at a Distance: What Is Transmitted and What Is Grown (preprint, Zenodo).

Science fiction has imagined teleportation in many ways. A person steps into a booth, dissolves, and is reassembled on another planet. A person is scanned, sent as a signal, and rebuilt from local matter. A person’s mind is uploaded into a computer and sent to the stars on a laser beam. All of these versions share one assumption: what arrives at the other end is the very same person, instantly and without loss. That assumption is exactly what makes their teleportation a utopia. A person cannot be reduced to data that can be captured, transmitted, and unpacked again. Part of a person can be transmitted: memory, knowledge, skills. Another part cannot: the living structure that binds all of this together. It can only be grown, slowly, and a little differently each time. So real teleportation, if it is possible, looks different from the science-fiction kind. What arrives is not the same person but a successor. How this could work, and why this version is not worse but more honest, is the subject of this article.

1. How the question was posed before

Parfit: is it the same person? Philosophers have discussed teleportation for a long time. In Reasons and Persons (1984), the British philosopher Derek Parfit described a teleporter that scans a person on Earth, destroys the original, and assembles a copy on Mars from local matter. No matter travels; only information does. Parfit asked: is this the same person?

The question is philosophical and social rather than technical. The copy preserves psychological continuity: memories, intentions, character. Imagine that the copy on Mars carries on your projects, loves the people you love, remembers your life, and pursues your plans as its own. Even if it cannot be proved that this is you in the strict sense, perhaps everything that matters for your survival has been preserved? Parfit himself leaned towards this view. In his thought experiment the original is destroyed. But Parfit also considered the case in which the original survives: then there are two identical people, one on Earth and one on Mars, and the question of which of them is the real one arises in full. His answer: what matters for survival is not identity but preserved psychological connectedness.

Transhumanists: mind instead of body. Supporters of transhumanism — the movement that seeks to overcome human biological limits through technology — went further. They combined two ideas: upload the human mind into a computer, and send that computer to the stars instead of a person. The advantage is obvious: a ship no longer has to carry a living body, a life-support system, and food for decades or centuries. It carries data and computing hardware, and the mind could even be sent by radio.

In Mind Children (1988), the American roboticist Hans Moravec described how the brain could be replaced, cell by cell, with a computer system. Later interstellar mission studies treat such mind transfer as the main saving in mass (Hein, 2014).

Three different claims need to be kept apart here:

1. The brain can be described precisely as information.

2. From this description one can build a digital system that reproduces the brain’s operation.

3. That system will not merely behave like the person but will actually have their consciousness and continue their personality.

The first two are scientific and engineering problems. The third is an unsolved philosophical problem, and everything else rests on it.

As dreams, the ideas of Parfit and Moravec are interesting: they give food for thought and a push for further ideas. But as working concepts they do not survive careful analysis, and they most likely belong among the utopias. Why this is so is explained by the theory on which this article rests.

2. The theory in brief

The theory of Causa Sui is set out in three books. You do not need to know them to read this article: everything required is given below.

A subject is one who acts on its own, answers for the consequences, fears error, and is able to come up with something new. A human being is a subject. The theory asks what a subject can arise from.

The main condition is inseparability. In a computer, program and hardware are different things: a program can be copied, moved to another machine, saved, and restarted. A living being does not work like that. Take the ability to ride a bicycle. It cannot be passed to another person like a file: you can explain how to keep your balance, but they will have to learn it themselves, with their own body and their own falls. The skill does not sit in the head separately from the one who has it. The theory calls this the inseparability of the program from its carrier and adopts it as its first postulate — a starting proposition that is not proved but laid down as a foundation, like the axioms of geometry.

Inseparability implies a prohibition: a subject cannot be “saved” and later “restored from the save” the way a computer can. In the theory this operation is called restore. If it can be done, what we have is not a subject.

Hence the answer to the transhumanists. A digital carrier is built precisely so that restore is possible for it. Therefore it cannot be a subject, and a mind placed in it will not be a copy of the person, even with generous allowances.

Two parts of a subject. To understand what can be transmitted after all, two parts must be distinguished.

• Matrices are the content: what the subject knows, remembers, and can do. Matrices may be living, but they may also be made of silicon. They can be captured quickly, the way a copy of a computer’s memory is taken. But such a copy is dead by itself: it is a set of records with nothing to use them and no way to use them.

• The graph is the living structure that connects the matrices and determines how the subject passes from one state to another, how it responds and decides. The word comes from mathematics, and there is a resemblance: each vertex corresponds to a particular state. But there the resemblance ends. A mathematical graph is a passive scheme: it only describes events that occur outside it. Our graph is an actor: it moves from vertex to vertex by itself and draws the information for this from the matrices of its vertices (Book 3, Ch. 27). The graph is not loaded; it grows, the way connections between nerve cells grow. In principle it can be read, but slowly. And while the reading is going on, the graph keeps living and changing, so the snapshot always describes its past, not its present.

Back to the bicycle: the knowledge that you turn the handlebars towards the side you are falling to is a matrix; it can be written down and passed on. The skill of keeping your balance itself is the graph; it can only be grown.

Successor and drift. In the second book the theory examines how one subject can continue in another. A new subject that has received the content of an old one is not the same subject but its successor. How much it differs from its predecessor the theory calls drift. Drift is never zero.

3. What we missed

All the parts of the answer were already in the project’s books. The first book examines two ways in which a subject can create its like. In the slow way, a new subject is grown from scratch: a living graph and starting matrices, after which it gains experience on its own. In the fast way, an already grown new subject immediately receives the whole content of the previous one. “Fast” is relative here, compared with the slow way. The same book explains why the best field for a subject on a living carrier is deep space: communication with Earth takes too long to control every step, and it is cheaper to create new subjects on the spot from local resources than to ship them from Earth. The second book introduced successors and drift; the third, the division into matrices and graph.

But the first book silently assumed that the old and the new subject are next to each other. The idea that content can be sent over a communication channel while the carrier is grown on the spot, or one grown in advance is used, was never stated. This article fills that gap; the second edition of the first book (Ch. 10) notes it briefly, with a reference to this article. It introduces no new propositions: everything said below follows from those already adopted.

4. What travels and what grows

Hence the main conclusion of the article. If matrices are transferable and the graph is grown, then distance is no obstacle to succession. Matrices travel over the channel at the speed of light. The graph need not be transported: it grows on the spot, or is already waiting there, grown in advance. The first postulate ceases to be merely a prohibition. It shows what in a subject can travel at all, and therefore in what sense a subject can be teleported.

The whole process falls into two stages of very different speed.

Stage one: growing the graph. It is slow, even if growth is artificially accelerated. Growth is governed by a growth program: it specifies how many connections, and of what kind, the graph should form with the matrices. The program itself is an ordinary description, and it can be sent over the channel in advance. But growth takes place in a biological medium, where some things turn out better and some worse. Already at this stage the grown graph differs from the ideal written in the program.

The graph does not have to be grown at the very moment of transfer. It can be grown in advance from the same program: then the receiver is already waiting, and the slowest stage is taken out of the transfer. Otherwise every transfer would stretch over the whole time of growth. Such a receiver accepts content well and quickly: its graph has been grown according to the scheme, and its matrices are almost empty (Sec. 6).

Stage two: transfer. Once the graph has grown on the spot, only the content is transmitted. Transmission over the channel runs at the speed of light. Writing into a prepared graph is fast too: the connections have already been grown and only need to be stimulated (see below). It can be repeated: experience can be transferred into the same grown graph again and again, as the original subject accumulates more.

This also shows why transfer is possible at all. Source and receiver are grown from the same program; their graphs have the same design, and the content of one fits the graph of the other.

How exactly can content be written into a living graph? There are two ways. The slow one applies when the graph is not prepared for writing: then all that remains is to guide its growth. This process has long been known in nature: it is learning. Writing according to a pattern taken from another subject is the same learning, only from a very detailed textbook, and it is not fast. The fast one applies when the graph has been prepared in advance: under the control of the growth program all its connections to the matrices have already been grown, but the matrices themselves are still empty. Then nothing needs to grow. It is enough to stimulate the ready graph, and it will write the needed content into its matrices in full by itself. It also matters who does the writing. According to the theory, only the subject itself writes into its own memory (Book 3, Ch. 27): this follows directly from the first postulate, since the subject’s history is part of the subject. Both ways keep this rule: the receiver writes by itself — with its growth or with its ready connections — and the pattern only guides or triggers the writing. Loading from outside would break this rule.

5. Where the discrepancy comes from

A successor never coincides with its source completely. The discrepancy in transfer is made up of three parts:

discrepancy ≈ difference of graphs + reading error + writing error

• Difference of graphs. Both graphs deviate from the growth program: the source, too, once grew imperfectly. So what matters is not how far the receiver departs from the program but how much the two real graphs differ from each other. The source’s content is “laid out” over its own graph but lands on another one that is almost the same. Almost — because both were grown from one program, and the difference between them is small. It is rather like the same piece played on two violins made by one maker from one drawing: the music is the same, the shades slightly different. If the difference is large, that is a defect of growth, and such a receiver is unfit for transfer. The difference accumulates slowly, during growth, and shows itself at the moment of transfer.

• Reading error: the inaccuracy with which the source’s content was captured.

• Writing error: the inaccuracy with which it was written into the receiver.

The sum is approximate: the parts can affect one another. For example, writing into a graph that differs strongly from the original will most likely be less accurate. This discrepancy is the successor’s drift at the moment it comes into being.

Moreover, the receiver goes on living its own life while it is being written into. The writing becomes an event in its own history. The successor ends up with two pasts: its own and the written one.

6. Why this does not contradict the theory

A natural objection arises: if a subject can be read and written, haven’t we arrived at the very restore that the theory forbids?

No. Restore is the return of the same subject to a previous state. Something else happens here: the content of one graph is laid onto another graph. Even with perfect reading and perfect writing, the difference of graphs will remain, because living things never repeat their growth exactly. This is another starting proposition of the theory (the third postulate): two specimens of the same design grown under the same conditions still turn out different. In the same way, identical twins with the same genes grow up to be different people. That is why an exact copy of a living carrier is impossible; only a near-exact one is possible (Book 3, Ch. 24, 35).

So the ban on restore rests here not on the imperfection of equipment, which may someday be fixed, but on the nature of living growth. This limitation is natural, and it is stronger than any technical one.

Along the way, the paradox of two instances that Parfit discussed disappears. If the original is kept, which of the two is “I”? For those who regard a person as nothing but information (in philosophy they are called functionalists), copy and original are one and the same, and the question is agonizing. In our theory it does not arise: these are two different subjects linked by succession. The original does not have to perish. The successor is neither an impostor nor a ghost but a continuer, and drift honestly shows how much it differs.

There is also a limitation. A prepared subject, whose everything has been grown according to the scheme and whose matrices are almost empty, accepts new content well and quickly. The opposite case is a subject whose matrices are already filled with its own experience. To write the new, the old has to be displaced, and anything can happen here, up to the subject’s refusal to change what it already has: it has its own boundary between “self” and “non-self”, like the immune system. Such a refusal is not a malfunction but a defence of integrity: the subject’s history is part of the subject, it has something to defend, and it defends itself. That is why it is better to write into a prepared carrier with almost empty matrices — such as a receiver grown in advance.

7. What has already been done in laboratories

Everything above was built as theory. But individual links of the scheme have already been obtained in experiments.

Writing a memory into a living brain without experience. Paper: G. Vetere, P. W. Frankland, et al., “Memory formation in the absence of experience,” Nature Neuroscience, 2019, vol. 22, pp. 933–940; open access. Researchers in Toronto created a fully artificial memory in mice. The odor was replaced with light stimulation (optogenetics) of a specific site in the olfactory bulb, and reward or punishment with stimulation of inputs to the brain’s reward center. Afterwards the mice began to avoid or seek out the real odor, although they had never encountered it together with punishment or reward. The artificial memory worked through the same brain circuits as a real one. Two things matter for us: a ready-made living network can be written into, and this is done not by loading data but through the brain’s own circuits.

Transfer between two brains at a distance. Paper: M. Pais-Vieira, M. A. L. Nicolelis, et al., “A Brain-to-Brain Interface for Real-Time Sharing of Sensorimotor Information,” Scientific Reports, 2013, vol. 3, article 1319; open access. Miguel Nicolelis’s group at Duke University linked the brains of two rats. While one rat (the “encoder”) solved a task, signals from its cortex were sent to the cortex of the other (the “decoder”), which learned to make the same choice relying only on those signals. In one experiment the rats were on different continents: in the USA and in Brazil. The decoder’s accuracy reached about 70%, against a theoretical ceiling of about 78%. These are reading and writing errors in pure form, and measured ones. This is not a measure of personal drift — what was transmitted was not a personality but a single skill — but it is a good analogy: even the skill did not reach the receiver in full. Out of a hundred attempts, the receiver repeated the source’s choice in only seventy. This loss is the discrepancy between them: part of it arises in writing (even with direct stimulation of the decoder’s brain, accuracy did not rise above 78%), part in reading. The theory predicts that such a discrepancy will not vanish however perfect the equipment becomes. One more detail: the decoder’s cortex began to represent not only its own whiskers but the encoder’s as well. The writing physically rebuilt the receiver’s brain.

We must say honestly what these experiments do not prove. They show that individual links are feasible. They do not settle the dispute over whether a person reduces to information: a functionalist will read them as confirming his position — information was transmitted, so information is all there is. What speaks in favor of our theory are the details: writing proceeds through learning, not loading, and the living tissue of the receiver is rebuilt.

8. What this changes for deep space

Deep space is the natural field for a subject on a living carrier. Transfer at a distance makes this picture considerably more realistic.

• What flies is a seed, not a subject. The growth program is sent to a distant planet, along with the little that is needed to grow a carrier from local resources. The subject’s content arrives later as a signal. And if the receiver has been grown in advance, nothing needs to fly at all: only the signal remains.

• The bottleneck moves. The slowest and most expensive part of settlement becomes not delivery but growing the graph on the spot. The pace of settlement is set by biology, not by the rocket. Receivers grown in advance take growth out of the transfer itself, but not out of settlement: they still have to be grown.

• Corrections can be sent later. If the subject on a distant planet begins to err in its values, corrections can be sent after it, even if years late. But there is a flip side: the original subject’s errors are transmitted ready-made along with the content.

• The channel works both ways. Experience gained on a distant planet can be returned the same way. The result is not one-way settlement but exchange between worlds.

• The main asset is reproducible growth. The more precise the growth program and the more stable the medium, the smaller the difference of graphs and the closer the successor is to the source. What is most valuable in such a system is not a powerful transmitter but reliable growing.

The limitations must be named too. Over interstellar distances transmission will take years. The channel is noisy, and noise will add its share of discrepancy. Yet according to the theory this is not only a defect: noise is one more source of that irreducible dissimilarity without which there is no subject.

Conclusion

Let us sum up the ground we have covered.

Teleportation as science fiction described it is impossible. The very same person cannot instantly appear in another place — neither in a booth that takes the body apart and reassembles it, nor in a computer sent to the stars. The reason is not the weakness of present-day technology but the way living things are built. A subject is not a file: its ability to live, decide, and answer for its decisions does not lie apart from the tissue in which it grew. That is why it can be neither saved nor sent in its entirety.

But there is also something in a subject that can be sent: its content — knowledge, memory, skills — and the program by which its living basis is grown. From this another kind of teleportation emerges, more modest and more honest. On a distant planet a living carrier is grown from the program that was sent — in advance or at the time of transfer. Then the content of the original subject is transferred into it over a communication channel. What arrives is not the same person but a successor: it remembers what its predecessor remembered and can do what the predecessor could do, yet it remains a separate living being with its own growth and its own fate.

How much the successor will differ can be stated in advance: the difference between the grown bases plus the errors of reading and writing. This discrepancy cannot be guaranteed to be zero, and it need not be: it is precisely what distinguishes a living successor from a dead copy. What matters is that it is measurable, and the first measurements have already been made in laboratories on mice and rats.

This path takes nothing away from the original person. They do not have to perish for a successor to appear, as in Parfit’s thought experiment. Both live, each has a life of their own, and between them there is a bond of succession, like that between teacher and pupil, only far closer.

For the exploration of deep space this is enough: instead of ships with people, one can send seeds, and send people as a signal. For philosophy it is more than enough: the old question “is it the same person?” ceases to be a paradox. The answer is “no, it is a successor,” and how close it is to the original is a quantity that can be estimated and measured.

About the role of AI

This article was written in dialogue with the language model Claude (Anthropic). The occasion, the key ideas, the two-stage scheme, and the two modes of writing belong to the author; the wording, the literature search, and the checking of the logic were joint work.

References

Hein, A. M. (2014). Transcendence Going Interstellar: How the Singularity Might Revolutionize Interstellar Travel. Centauri Dreams, 13 June 2014.

Moravec, H. (1988). Mind Children: The Future of Robot and Human Intelligence. Cambridge, MA: Harvard University Press.

Pais-Vieira, M., Lebedev, M., Kunicki, C., Wang, J., & Nicolelis, M. A. L. (2013). A Brain-to-Brain Interface for Real-Time Sharing of Sensorimotor Information. Scientific Reports, 3, 1319.

Parfit, D. (1984). Reasons and Persons. Oxford: Clarendon Press.

Shakhanov, I. (2026). Project Causa Sui: A Minimal Architectural Theory of Subjecthood. Preprint, Zenodo. Concept DOI 10.5281/zenodo.21460340.

Shakhanov, I. (2026). Project Causa Sui: A Carrier Criterion for Subjecthood, Applied to Transhumanism. Preprint, Zenodo. Concept DOI 10.5281/zenodo.21641230.

Shakhanov, I. (2026). Project Causa Sui, Paths of Implementation — The Mechanism of the Subject, a Simulated World, and a Falsifiable Prediction. Preprint, Zenodo. Concept DOI 10.5281/zenodo.23243298.

Shakhanov, I. (2026). Succession of a Subject at a Distance: What Is Transmitted and What Is Grown. Preprint, Zenodo. Concept DOI 10.5281/zenodo.23685278.

Vetere, G., Tran, L. M., Moberg, S., Steadman, P. E., Restivo, L., Morrison, F. G., Ressler, K. J., Josselyn, S. A., & Frankland, P. W. (2019). Memory formation in the absence of experience. Nature Neuroscience, 22, 933–940.