The Feeling of Life Itself: Why Consciousness Is Widespread but Can't Be Computed

by Christof Koch

Koch describes how the theory explains many facts about the neurology of consciousness and how it has been used to build a clinically useful consciousness meter. The theory predicts that many, and perhaps all, animals experience the sights and sounds of life; consciousness is much more widespread than conventionally assumed. Contrary to received wisdom, however, Koch argues that programmable computers will not have consciousness. Even a perfect software model of the brain is not conscious. Its simulation is fake consciousness. Consciousness is not a special type of computation--it is not a clever hack. Consciousness is about being

Genres: Science, Nonfiction, Philosophy, Neuroscience, Psychology, Artificial Intelligence, Brain

280 pages, Kindle Edition

Published October 1, 2019

About the author

is an American neuroscientist best known for his work on the neural bases of consciousness. He is the President and Chief Scientific Officer of the Allen Institute for Brain Science in Seattle. From 1986 until 2013, he was a professor at the California Institute of Technology.

Reviews

[Katia N · Rating 5 out of 5]

It has been 3 weeks since I’ve written any review of sensible length due to problems both with hardware (my pc is being constantly in use) and software (my mind is being constantly in use as well). This is because it has been almost 3 weeks since my son has been to his normal school. They had a covid case in his year group and he has been sent home. But some students from his year have been told to stay. And, just when the first batch was about to come back they had another case. All went in circle again. When I tried to raise the question of mask wearing during the lessons in the classroom of 25 teenagers sitting closely together, other parents said it was unnecessary as their children would be “uncomfortable.” End of story. Let’s see what the next week brings. For now, when I was not involved in recalling what were enzymes or covalence as I was often recruited for help and motivation, I was not allowed near my Mac in any case. But here we go. Today is the first day I am “allocated” a bit of time for my own business.

This book. I am a bit partial to Koch as a scientist, exciting storyteller and very decent human being. I also find the theory that he supports very elegant and intuitively appealing. So my review might come across as a bit biased. But one cannot get rid of subjectivity. Especially while we talk about consciousness. I’ve read his previous book as well. And while that one was more personal, this one is probably better in summarising the current state of science in this area.

Koch’s definition of consciousness is a broadly accepted one I guess. It is any subjective experience from pain to taste, emotions, colour perceptions and more sophisticated advanced matters like feeling of self (the last however is not necessary to be conscious). I’ve recently read much older and more famous book The Origin of Consciousness in the Breakdown of the Bicameral Mind. There, consciousness is defined more like ability of introspection directly related to language and self-awareness. For me, it was too narrow a definition for an experience. That definition would also exclude every other living creature. I found Koch’s approach more constructive. In fairness, I think the majority of scientists dealing with the subject would agree with Koch’s broad definition now. Janes's book has been written in the 70s.

Koch is a big propagator of the Integrated Information Theory (IIT) designed by the fellow neuroscientist Guilio Tononi. The book discusses this theory, its implications and touches upon other modern theories in the area.

IIT starts from defining the axioms what constitutes an experience. There are 5 of them in total. If someone is properly interested in the theory, I would post a link to the article by Tononi where all of this is explained brilliantly. For my purpose here, I would just say that each experience generates something different from anything else including any other possible experiences. It sounds obvious as any axiom should. But then IIT exploits on which possible substrate such a phenomena could be generated. The answer is only on a physical system which is very highly integrated with lots of feedback loops. Such system would be able “feel like” something from inside. While other systems, less integrated with more feedforward architecture or the modules would only be able to feel like the lowest module with the highest integration.

The implication of that is that we and other creatures with brains or sophisticated nervous systems would be sentient; an electrical circuit with feedback loop would possess some very minimal intrinsic perspective. It would be the case for an atom or a cell. But no spoons, tables or computers would be able to be conscious. In the case of a computer, theoretically it is possible. But modern computers are not integrated enough and could not be made to feel anything unless they are sort of a hybrid with a human (“losely mimic neuronal architectures”). As it stands, they could easily simulate anything, even consciousness. But they would not be able to know “what it feels like” to be. As Koch says, there is a big difference between “being” and “doing”.

“Conventional digital computers, built out of circuit components with sparse connectivity and little overlap among their inputs and their outputs, do not constitute a Whole. Computers have only a tiny amount of highly fragmented intrinsic cause-effect power, no matter what software they are executing and no matter their computational power. Androids, if their physical circuitry is anything like today’s CPUs, cannot dream of electric sheep.”

This theory is in contradiction with the currently prevailing orthodoxy that “Mind-as-software is an unspoken background assumption that needs no justification.” For Koch, “It is as obvious as the existence of the devil used to be.”

This prevailing view in neuroscience and Silicon Valley is called computationism. It more or less postulates that consciousness is the result of a software run on our brain. It cannot exist without the input from outside and serves to produce the output which would go outside as well. The best theory from that camp is Global Workspace Theory (GWT): 


“When activity in sensory cortices exceeds a threshold it triggers a global ignition, whereby information enters the global neuronal workspace. This information then becomes available to a host of subsidiary processes such as working memory, language, planning, and voluntary action. The act of globally broadcasting this information is what makes this data conscious. Nothing more and nothing less. Our stance is based on a simple hypothesis: What we call “consciousness” results from specific types of information-processing computations, physically realized by the hardware of the brain.”

Koch quotes here the GWT's the one of the main proponents, Stanislas Dehaene.

Interestingly, there is currently something called “adversarial collaboration” between IIT and GWT. By the end of 2021 we should find out who is right. Where there is something we intrinsically possess or we are just processing machines like computers. The details of this collaboration is in the link below.

Coming back to the IIT, another implication of it that only the module with the highest integration would have consciousness. In other words, if we would be able to establish a strong hypothetical bridge between two independent brains, they would cease to feel like independent persons. And the combined brain would feel like one instead. So the one for the price of two effectively.

However, Koch argues that our metaphors about feeling like a country or a corporation are just that, metaphors:

“Thus, the Anima Mundi or world soul is ruled out, as it requires that the mind of all sentient beings be extinguished in favour of the all-encompassing soul. Likewise, it does not feel like anything to be the three hundred million citizens of the United States of America. As an entity, the United States has considerable extrinsic causal powers, such as the power to execute its citizens or start a war. But the country does not have maximally irreducible intrinsic cause-effect power. Countries, corporations, and other group agents exist as powerful military, economic, financial, legal, and cultural entities. They are aggregates but not Wholes. They have no phenomenal reality and no intrinsic causal power.”


This is another thing I really like about this the IIT. But lets hope that internet would not connect us to a degree that we would stop feeling like anything and instead would become the submodules of the one big Whole that would be the only one having a consciousness.

There is a lot of interesting stuff to think about in the book. Koch talks about some practical implications which are being used already. For example they built up a machine which helps to identify consciousness in a severely disabled people who cannot communicate.

But the book ends with Koch’s thoughts not about science, but about moral implications of this theory. The two main ones, that the animal’s rights need to be much better protected as they are sentient beings. And another one that the intelligence should not be confused with consciousness.


“Finally, if we link moral rights to certain functional abilities, such as imagination or intelligence, than we will sooner or later have to admit software running on digital computers into the club. What happens if they become capable of cognitive feats people cannot match? They would then leave us morally in the dust, even though, per IIT, they would not feel anything.”

Links:

Adversarial colalboration:
https://www.templetonworldcharity.org...

Guilio Tononi on IIT:
https://royalsocietypublishing.org/do...

[Alan Green · Rating 4 out of 5]

Koch is a neuroscientist, and the book is devoted to promoting a theory of consciousness, which in its essentials he attributes to his mentor Crick (of Crick and Watson,) and to G. Tononi. The core problem addressed by Koch is what philosophers call "the hard problem" - how does being a subject of experience with a rich and logically private inner life emerge from a bunch of wet stuff (your brain)? We can infer (or to use Koch's technical expression "abduct") that other persons have experiences in some way similar to our own, but we can never actually observe the inner experiences of another, let alone explain how neurological circuits, however complex, give rise to experience. The book has broadly two aspects: a survey of the philosophical literature and fascinating brain science relevant to the problem; and a polemic consisting of his own elaborate thesis: "integrated information theory" or "ITT." The former aspect is excellent; but the latter, although extremely interesting is a weak sustained non sequitur. No one however should be disappointed that Koch has failed to solve the hard problem, in relation to which in the views of many no fundamental progress has been made in 2,500 years since Aristotle first addressed it.

There is much fascinating and detailed material about the brain science of human consciousness and its "neural correlates". It is fascinating, to give one example, that on Koch's account the prefrontal cortex is not implicated in consciousness. Some of the work surveyed has clinical implications, for example deciding whether patients apparently in a persistent vegetative state are conscious or not. It is frightening that the evidence strongly suggests that a significant cohort of such patients are in fact conscious. There is also a fascinating discussion of the postulated consciousness of animals, other living beings, and even inanimate complex structures. It is one matter to accept, as most of us do that one's pet dog may be conscious. The evidence and argument that consciousness may descend the evolutionary tree to worms (Darwin's view) and even bacteria, is apt to be disturbing, and threatens overwhelming moral dilemmas.

The core theory of the book is of course "ITT". It seems to me that the basis of this theory is an extended analogy derived from certain characteristics consciousness seems to have: that it is structured, informative, integrated, and definite (and also has a point of view and takes place through time.) From this Koch moves to an abstract system consisting of logic gates that he claims have the same characteristics. I cannot understand the detail of the argument here, which depends on a knowledge of aspects of computer science not explained by Koch. However, that the argument is based on an analogy seems clear. An essential feature of such a system is self-causality. Koch loses me when he claims that such a system is strictly identical to being conscious - a large claim when one consider Leibniz's law of identity (how does it really feel good or bad for this system?) It may be a reasonable speculation from the brain science that the nature of the neural correlates of consciousness are similar systems; but this does not solve the hard problem of how it happens that such systems are subjects of experience. Finding neural correlates does not explain the how. To this Koch's response is: well that's just the way it is, which he compares to similar foundational truths in physics. (There is a point of view here as to where it is likely that explanation comes to a satisfactory end.) Finally, I am not convinced that Koch is right to dismiss the possibilities that areas of the brain such as the frontal cortex, the amygdala, and the brain stem, do not have experiences which "we" do not have access to. (These are certainly difficult issues to attain clarity about.) In relation to these, and other issues (such as the possibility of panpsychism) the "argument" often consists of mere assertion, and weak unsupported claims of the plausibility of Koch's point of view.

A claimed consequence of Koch's theory is that current standard computers ("Von Neumann machines") could never be conscious, because their architecture precludes self-causing causal feedback loops of the right type.

In summary: an excellent book with much fascinating material, and none the worse for promoting a thought provoking theory.

[Markus · Rating 5 out of 5]

Bewusstsein und Informatik - endlich einmal ein Sachbuch, das mich vollends begeistert hat, nicht zuletzt angesichts des Schwurbels, den so manch prominenter Philosoph (aber auch IT-Gläubige) zu dem Thema verbreiten!
Es geht um die von Giulio Tononi und Kollegen entwickelte Integrated Information Theory, kurz IIT, eine der interessantesten Theorien zur Erklärung des Bewusstseins. Vielleicht sind die Wissenschaftler hier einer großen Sache auf der Spur. Auch die Relativitätstheorie war vor 100 Jahren "nur" eine Theorie, die sich erst bewähren musste und die in der Folge unser Verständnis von Raum und Zeit radikal erneuert hat.

Der Autor des Buchs, Christof Koch, ist einer der führenden Bewusstseinsforscher. Er studierte Physik und Philosophie in Tübingen und forschte zusammen mit Francis Crick¹ (ja genau der) am neuronalen Korrelat des Bewusstseins. Heute arbeitet Koch am Allen Institute for Brain Science in Seattle.

Anders als die konkurrierenden Theorien² des Bewusstseins geht die IIT vom Phänomen des Erlebens aus und schreitet von dort zu dessen Substrat, dem Gehirn fort. Um Bewusstsein zu ermöglichen, setzt sie unter anderem eine intrinsische Kausalkraft voraus. Bewusstsein ist demnach eine fundamentale Eigenschaft von allem, das sich selbst ursächlich beeinflussen kann.


[Ernst Mach - Innenperspektive]

Diese Kausalkraft lässt sich für jedes System berechnen oder messen, egal ob es sich um einen Neuronenverband im Gehirn oder eine Anordnung von Logikgattern aus der Informatik handelt. Ich muss gestehen, dass ich Kapitel 8, wo es am Beispiel eines einfachen Schaltkreises ins Eingemachte geht, mehrfach lesen musste, um den Sachverhalt zu erfassen. Meine Faszination wuchs mit dem Verständnis, aber ich will hier niemanden langweilen. Für Unerschrockene verweise ich hier nur auf integratedinformationtheory.org und auf scholarpedia.org, eine gut aufbereitete Übersicht der IIT.
Dankenswerterweise gibt es auch eine freie Software als Pythonmodul zum Download, damit auch unsereins damit herumspielen kann :)) Es soll jedoch nicht verschwiegen werden, dass die Berechnung der Integrierten Information des Fadenwurms, C. elegans, mit seinen 302 Neuronen heutige Computer noch überfordert.

Jede Theorie beinhaltet eine Reihe an Voraussagen und Konsequenzen, die sich in der Wirklichkeit bewahrheiten müssen und so die Theorie bestätigen oder eben nicht. Es ist auf Basis der IIT bereits gelungen, den Protypen eines Messgeräts für Bewusstsein zu bauen. In der klinischen Praxis wäre so ein Gerät ein Segen, man denke an das Dilemma von Komapatienten, von denen man nicht weiss, ob sie bewusst sind, wenn sie körperlich nicht in der Lage sind, die geringste Reaktion zu zeigen. Erste Versuche verliefen vielversprechend. Bei allen Patienten, von denen man weiss, dass sie bewusst sind, (zB Locked-In-Syndrom) zeigte das Gerät Bewusstsein an, bei den meisten Patienten, von denen man annimmt, dass sie nicht bewusst sind, zeigte das Gerät auch nichts an, aber eben nicht bei allen. Es ist beunruhigend, dass laut Messung einige dieser Komapatienten trotzdem bewusst zu sein scheinen.

Die IIT trifft auch recht ungewöhnliche und kontraintuitive Voraussagen. Eine davon betrifft den geteilten Geist von Splitbrain-Patienten³. Die Theorie unterstützt die Annahme, dass jede der beiden Gehirnhälften tatsächlich über ein eigenes Bewusstsein verfügt. Eine andere, für mich schwer zu akzeptierende Konsequenz wäre, dass auch ein bestimmter Schaltkreis aus Logikgattern und lichtempfindlichen Dioden ein Minimum an Bewusstsein haben müsste.

Mehr als spannend ist eine Diskussion im Internet zwischen Tononi und dem Quantenphysiker Scott Aaronson, dessen Intuition sich dagegen sträubt, dass ein Netzwerk, dessen Knoten alle auf OFF stehen, das also nichts tut, bewusst sein soll. Offenbar sind wir so vom Tun vereinnahmt, dass wir das reine Sein nicht mehr wahrhaben wollen. Tononis ausführliche Antwort beginnt mit: Scott solle doch eine weisse Wand betrachten ... (jeder Zenmeister weiss, was leeres Bewusstsein, reines Gewahrsein bedeutet).

Eine weitere Konsequenz der IIT wäre es, dass Bewusstsein in kontinuierlichen Graden auftritt. Lebewesen können danach mehr oder weniger bewusst sein, von knapp über Null ansteigend auf einer nach oben offenen Skala. Mit der Konsequenz, dass auch Einzeller oder Bakterien irgendeine Form von Erleben hätten, wenn auch ein sehr geringes. Dass Menschen, die langfristig und regelmäßig meditieren, einen besonders hohen Grad an Bewusstsein aufweisen, was in zahlreichen neurologischen Studien schon länger angenommen wird, würde sich bestätigen. Die nächsten Jahre der Forschung dürften jedenfalls sehr spannend werden.

Wenn man den Grad von Bewusstsein berechnen kann, dann folgt die Frage nach seiner Digitalisierung auf den Fuß und die Frage ist angesichts der rasanten Entwicklung der Technologie auch mehr als berechtigt. Wird unser Mobiltelefon in Zukunft bewusst sein oder werden unsere Enkel ihren Geist in die Cloud hochladen? Damit befassen sich die weiteren Kapitel des Buchs. Die entsprechenden Schlussfolgerungen der IIT sind für mich beruhigend und bestätigen mein Bauchgefühl.

Haben Sie sich jemals gefragt, warum Astrophysiker, dıe Schwarze Löcher simulieren, nicht in ihre Supercomputer hineingesaugt werden? Wenn ihre Modelle die Wirklichkeit so genau nachbilden, warum schließt sich die Raumzeit dann nicht um den Computer, der die Simulation ausführt, und erzeugt ein kleines Schwarzes Loch, das den Computer und alles um ihn herum verschluckt? Weil Gravitation keine Berechnung ist! Gravitation hat echte extrinsische kausale Kraft. Diese Kräfte können funktionell simuliert werden [...], aber das verleiht diesen Simulationen keine kausale Kraft.

Wunderbar erklärt! Simulation ist nicht Realität. Die integrierte Information von Computerhardware, ihre intrinsische Kausalkraft liegt nahe bei Null. Wir können mit Computern alles mögliche simulieren, irgendwann auch Bewusstsein - doch simuliertes Bewusstsein ist Fake-Bewusstsein. Um eine wirklich bewusste Maschine zu bauen, würden wir neuromorphe Hardware brauchen, was derzeit nicht absehbar ist⁴.

In den letzten Jahrzehnten hat sich mit Unterstützung der HiTech-Industrie und der philosophischen Fakultäten vor allem des angelsächsischen Raums eine Erzählung durchgesetzt, die den menschlichen Geist als Software beschreibt, die auf einem feuchten Computer, unserem Gehirn läuft: Der Mensch als Turing Maschine. Sensorischer Input wird gespeichert, in mehreren Schichten verarbeitet, daraus folgt motorischer Output in Form von Bewegung, Handlung, Sprache. Der amerikanische Philosoph Daniel Dennett meint überhaupt, Bewusstsein gäbe es gar nicht, es sei nur eine von der Evolution entwickelte Einbildung, um dem Menschen ein praktisches Userinterface zu geben.

Kochs Analyse und kritischer Kommentar zu dieser Entwicklung im letzten Teil des Buchs sind erfrischend eindeutig. Er stellt sich mit klaren Argumenten gegen diese Tendenzen und entlarvt sie als Mythos einer Kultur, die glaubt, in ihrer Rationalität über jeden Mythos erhaben zu sein! Zum Schluss betont Koch seine Überzeugung, dass jedes zelluläre Leben in irgendeiner Form empfindungsfähig und dass viel mehr Lebewesen leidensfähig sind, als man heute glaubt. In dieser Erkenntnis liegt implizit eine Verpflichtung zu ethischer Verantwortung gegenüber allen, aber ganz besonders den kleinsten und schwächsten Mitbewohnern dieses Planeten. Christof Koch bekennt sich deshalb als Vegetarier und als überzeugter Verfechter von Tierrechten. In seinem Schlusssatz nähert er sich der buddhistischen Sicht des Lebens an, womit er nicht nur meine Begeisterung, sondern auch meine Sympathie geweckt hat:

Die buddhistische Einstellung zur Empfindungsfähigkeit spiegelt diese Sichtweise wider. Wir sollten alle Tiere als bewusste Wesen behandeln, als ob sie fühlten, wie es ist zu sein. Das muss die Leitlinie sein, wie wir uns unseren Mitreisenden durch diesen Kosmos gegenüber verhalten, Wesen, die sich nicht gegen unsere Zäune, Käfige, Klingen, Kugeln und unser ständiges Streben nach Lebensraum zur Wehr setzen können. Eines Tages könnte die Menschheit danach beurteilt werden, wie sie ihre Verwandten im Lebensbaum behandelt hat. Wir sollten eine universelle ethische Haltung gegenüber allen Kreaturen einnehmen, ob sie nun sprechen, schreien, bellen, jaulen, heulen, zirpen, quieken, summen oder stumm sind. Denn alle erleben Leben, eingerahmt von zwei Ewigkeiten.

--

¹ Francis Crick, James Watson und Rosalind Franklin entdeckten 1953 die Struktur der DNS.

² Die beiden neben der IIT relevanten Modelle sind das
- "global workspace model" von Bernhard Baar inkl. der Weiterentwicklung von Stanislas Dehaene und
- "orchestrated objective reduction" (Orch OR) von Roger Penrose.

³ Es ist schon lange bekannt und ziemlich eigenartig, dass bei Menschen, denen aus therapeutischen Gründen das Corpus Callosum, die Verbindung der beiden Gehirnhälften chirurgisch durchtrennt wurde, jede Gehirnhälfte ihren eigenen Geist zu haben scheint. Weil das Sprachzentrum vorwiegend in einer Hälfte angesiedelt ist, ist das schwer zu überprüfen.

⁴ Cerebrale Organoide, aus Stammzellen gezüchtet, besitzen laut IIT geringes Bewusstsein, aber keinerlei Intelligenz. Aber wer weiss, was noch alles möglich sein wird.

[Benji · Rating 4 out of 5]

'Consciousness is not a clever algorithm. Its beating heart is causal power upon itself, not computation. And there's the rub: causal power, the ability to influence oneself or others, cannot be simulated. Not now, nor in the future. It has to be built into the physics of the system.'

[Cheenu · Rating 2 out of 5]

This book is a short primer on the Integrated Information Theory (IIT) of Consciousness.

This piqued my interest because I've seen this theory mentioned a few times - it seems to be one of the more popular theories of consciousness (there are literally more than you can count).

So the gist of the theory, as I understand from the book, is that there is a calculable "maximally irreducible cause-effect structure" (stay with me!) in every distinct entity.

What you do is take the current state of the system and math out how much it constrains the previous state (cause) and how much it constrains the future state (effect).

Whichever of these is higher is "cause-effect structure" denoted by Φ.

Stuff (such as brains) can contain many multiples of such systems (distinct regions such as hippocampus, amygdala etc) - which when integrated have a higher Φ than the sum of its constituent parts.

When the Φ of something higher order (such as a group of people) is not greater than the sum of its constituents (Φ of individual people), then there can be no consciousness attributed to the higher order system.

---

My thoughts?

It's an interesting theory but the point at which it fell through for me was when the author discusses "inactive vs inactivated neurons".

So you take some Buddhist monk at a very heightened state of mediation. And then you take someone who ate a pufferfish, which contains a neurotoxin that shuts down the synapses across the brain.

You stick EEGs on their heads and you find that the brain activity looks very similar (massively quietened).

However, one is experiencing a very heightened state of consciousness while the other is nonresponsive, literally unconsciousness.

What gives?

According to the author, the difference is that the neurons are still connected in the former case (even though not actively firing) but disabled in the latter case (so cannot fire).

This results in very different Φ values for both cases.

To me, this sounds logically equivalent to the concept of a "soul" - a metaphysical entity that exists independent of physiology.

[Oskari · Rating 2 out of 5]

Arvostelu venyi vahingossa yhtä pitkäksi kuin kirja... Tsemppiä.

Tietoisuus on lempiaiheitani sekä psykologiassa että filosofiassa ja luen mielelläni lähes kaikkea siihen liittyvää. Tästä huolimatta olen suhtautunut panpsykismiin (epäreilun) skeptisesti, sillä se on omassa pienessä päässäni yhdistynyt liian vahvasti New Age -sekamelskaan, jossa buddhalainen ja hindulainen filosofia, Jung, James, psykedeelikokemukset ja käpyrauhaset sulautuvat käsittämättömäksi sopaksi. Minulla on ollut tiedossa, että sillä on modernin tietoisuustutkimuksen piirissä muutama nimekäs kannattaja, mutta olen silti pitänyt sitä enemmänkin marginaalisena puuhasteluna kuin vakavasti otettavana suuntauksena. Kun tietoisuuden neurotieteen kurssilla sitten selvisi, että Giulio Tononin kehittämä Integrated Information Theory (IIT) panpsykistisine seurauksineen on viime vuosina noussut yhdeksi johtavista tietoisuusteorioista, päätin tarttua tähän ikuisuuksia lukulistalla lojuneeseen Kochin kirjaan. En valitettavasti voi sanoa vakuuttuneeni ja olen listannut tähän joitakin syitä. Olen myös yrittänyt tiivistää IIT:n lyhyesti alle, mutta se on teoriana aika monimutkainen ja jargonintäyteinen vyyhti, joten jos aihe ei ole etukäteen tuttu, arvostelusta ei välttämättä ole iloa.

IIT:n perusidea on tämä: kuvailemalla tietoisen kokemuksen ominaisuudet ("aksioomat") voidaan päätellä mitkä fyysiset ominaisuudet ("postulaatit") vaaditaan kokemiseen. Päättelyketju siis kulkee fenomenologiasta fyysiseen todellisuuteen. Skippaan pari vaihetta, mutta lopputuloksena IIT:n sisältämistä aksioomista ja postulaateista on se, että jokaisen tietoisen kokemuksen voidaan ajatella olevan identtinen ns. maksimaalisesti jakamattoman syy-seuraus-struktuurin kanssa (maximally irreducible cause-effect structure). Tämä sanahirviö tarkoittaa yksinkertaistettuna järjestelmää, jolla on kausaalista voimaa itseensä - sen tämänhetkinen tila johtaa sen tulevaan tilaan, ja vastaavasti sen tämänhetkinen tila on seurausta sen aiemmasta tilasta. Maksimaalinen jakamattomuus viittaa siihen, ettei kokonaisuutta voida enää jakaa pienempiin osiin muuttamatta järjestelmää toiseksi. Struktuurin (kuten hermosolujen muodostaman verkoston) tila tietyssä hetkessä määrittää kokemuksen sisällön, kvaliat. Struktuurin osille ja niiden suhteille voidaan myös suorittaa laskutoimitus, jolla saadaan selville sen jakamattomuus eli integroidun informaation määrä, fii. Fii on numeerinen arvio siitä, "kuinka paljon" jossakin järjestelmässä on tietoisuutta. Jokainen fysikaalinen järjestelmä, joka täyttää IIT:n määrittämät 5 postulaattia ja jonka fii on suurempi kuin nolla, on fenomenaalisesti tietoinen eli sen olemassaolo tuntuu joltakin sille itselleen.

Miksi sitten IIT:n seurauksia voi pitää panpsykistisinä? Jos tietoisuus määritellään integroiduksi informaatioksi, sitä on lähes kaikkialla. IIT, loogiseen päätepisteeseensä vietynä, tekee ameeboista, hehkulampuista ja ehkä jopa atomeista tietoisia. On ehkä älyllisesti epärehellistä hylätä jokin teoria siksi, että sen johtopäätökset ovat epäintuitiivisia, mutta minä en ainakaan toistaiseksi suostu hyväksymään sitä, että kattolamppuni voisivat tuntea jotain. IIT:n määrittämät tietoisen kokemuksen aksioomat ulottuvat periaatteessa kaikkeen tietoisuuteen, mutta ne on kehitetty meille ymmärrettävissä olevien kokemusten eli ihmisten ja joidenkin eläinten kokemusten pohjalta. Jos hehkulamppujen kaltaiset todella yksinkertaiset järjestelmät ovat tietoisia, on niiden kokemusmaailma meille niin mahdoton käsittää, että tietoisuus-sanan käyttäminen on joka tapauksessa vähintäänkin kyseenalaista.

Toinen, minulle suurempi ongelma IIT:ssä on sen suhde tietoisuuden vaikeaan ongelmaan. Koch väittää IIT:n kiertäneen ongelman tyydyttävästi: "IIT does not bang its head against this cement wall; it doesn't try to squeeze the juice of consciousness out of the brain. Rather, it starts with experience and asks how matter must be organized to support the mental." (s. 74) Mielestäni tämä on silmänkääntötemppu. IIT:n mukaan tietoisuus on integroitua informaatiota. Sisäinen kokemus ja maksimaalisesti jakamaton syy-seuraus-struktuuri ovat identtisiä, minkä vuoksi on järjetöntä kysyä, miksi tuolla struktuurilla on fenomenologinen puoli. Otetaan kuitenkin esimerkki ihmisaivoista: jakamattoman struktuurin tilan selvittämiseksi ja fiin laskemiseksi täytyy tietää jokainen tietoisen kokemuksen mahdollistava neuroni ja kaikkien näiden neuronien yhteydet toisiinsa. Konservatiivisesti arvioituna se tarkoittaa kymmeniä miljardeja neuroneja, joiden yhteyksien määrää en edes uskalla ajatella - joka tapauksessa nykyteknologialla mahdotonta monesta eri näkökulmasta. Kuvitellaan, että meillä on kaikki tämä tiedossa; osaamme kertoa täsmälleen, mitkä ruusun punaisuuden kokemuksen hermostolliset korrelaatit ovat. Mitä lisätietoa saadaan, kun määrittelemme verkoston tilan integroituna informaationa tai laskemme fiin? IIT:n mukaan tiedämme nyt, miksi meillä on tietoinen kokemus ruusun punaisuudesta, tai oikeastaan olemme määritelleet ruusun punaisuuden kokemuksen tietoisuuden olemuksesta käsin. Kochin vastalauseista huolimatta minä kysyn edelleen, että miksi integroitu informaatio tuntee mitään? IIT ei ratkaise tai edes kierrä vaikeaa ongelmaa, se vain hieman siirtää sitä eteenpäin.

Kirjan loppupuoli oli hyökkäys funktionalismia/komputationalismia vastaan. IIT:n mukaan nykyisellä tavalla kehitetyt tekoälyt eivät voi olla samalla tavalla tietoisia kuin ihmiset. Ne tekevät samoja asioita kuin ihmiset, mutta tämä on saavutettu hyvin erilaisella tavalla kuin aivoissa. Vaikka tietokoneiden pyörittämä softa simuloi ihmisten kognitiivisia prosesseja tehokkaasti, itse laitteiston maksimaalinen jakamattomuus on hyvin pieni: "Computers have only a tiny amount of highly fragmented intrinsic cause-effect power, no matter what software they are executing and no matter their computational power. Androids, if their physical circuitry is anything like today's CPUs, cannot dream of electric sheep." (s. 166) Vastaavasti tietoisuutta ei voi siirtää pilveen, vaikka aivojen konnektomi olisi täydellisesti kopioitu. En ole perehtynyt ohjelmointiin tai tekoälyyn niin hyvin, että voisin esittää tästä omia ajatuksia. Kuvittelisin, että IIT ei ole Piilaakson teknologiagurujen lempiteoria.

Itse uskon, että IIT on käyttökelpoisimmillaan viitekehyksenä NCC-tutkimuksissa (neural correlates of consciousness) eikä niinkään kaiken kattavana tietoisuusteoriana. IIT kuvaa aivojen toimintaa monella tapaa yhteneväisesti neurotieteellisten löydösten kanssa, joten tietoisuuteen vaadittavia hermoverkostoja voitaisiin mallintaa integroituna informaationa ilman sen suurempia ideoita kokemuksellisuuden luonteesta; esimerkiksi tietoisuuden tilan muutokset anestesian aikana voitaisiin ehkä selittää integraation hajoamisella. Tästä on jo olemassa joitakin alustavia tutkimuksia. Epäselvää on, että mitä lisäarvoa tästä on nykyisille menetelmille, ja että ovatko fiin epäsuorat ja helppokäyttöisemmät approksimaatiot kuten PCI (perturbational complexity index) toimivia mittareita.

Koch kirjoittaa IIT:stä todella ymmärrettävällä ja elegantilla tavalla, mutta suhtautuu lemmikkiteoriansa totuudenmukaisuuteen liian itsevarmasti ja muihin teorioihin alentuvasti. Välillä Kochin innostavat, karmivat ja toisinaan eroottiset ennustukset IIT:n käytännön sovelluksista menevät scifin puolelle, mutta toisaalta spekulaatio on aina jännittävää luettavaa. Kirja on suunnattu suurille yleisöille ja alussa selitetään pitkään mielenfilosofian ja tietoisuustutkimuksen perusasioita, mikä oli minulle pettymys, sillä olisin halunnut lukea IIT:n teoreettisesta perustasta yksityiskohtaisemmin. Täytyy joka tapauksessa vielä lukea Tononin kirjoituksia. IIT on omilla meriiteillään kiinnostava teoria, mutta minua mietityttää, että olisikohan tätä modernia panpsykismiä otettu yhtä vakavasti, ellei sillä olisi Christof Kochin kaltaista arvostettua tutkijaa näkyvimpänä kannattajanaan. Jos nimittäin Oskari Virtanen Turun yliopistosta julistaisi, että "kolibakteerit ovat tietoisia ja minä olen laskenut sen kaavalla", saattaisi jostain kaukaa välittömästi kuulla kaikuvan: "Vitun hipit!"

[Dmitry Khvatov · Rating 2 out of 5]

I found this book to be a disappointing read as it promoted several pseudoscientific concepts without any credible evidence. The author's claims lacked scientific rigor, and I was left feeling misled by the book's content. I would not recommend this book to anyone looking for reliable information on the topic.

[Ali · Rating 4 out of 5]

Koch is an engaging, provocative writer and thinker. This book was wildly different in tone to Confessions of a Romantic Reductionist, but both have provided a pretty rich intellectual feast. I got so frustrated trying to explain this book to my partner, that I got him to read a bit, and then he read most of it, and then we talked for days about all the things we disagreed with, and where Koch is not at all convincing and so on. Frankly, we're still going a bit.
Whereas in previous work, Koch defies the dispassionate scientist stereotype, letting his passion for understanding drive the narrative, the Feeling of Life Itself is an attempt to thoroughly explain Giulio Tononi's Integrated Information Theory of consciousness. It is much more of a traditionally structured science book than I expected, and there is little of the 'big issues' that Koch does not tackle, even if much of the time I found the argument unconvincing.
IIT is a theory which argues, effectively, that consciousness is produced by the interaction of binary neurons governed by both internal and external constraints. This interaction - mappable as a series of logic gates - creates a "Whole" which is 'irreducible' to individual components, as the system is causal upon itself. This provides a measurable approach to consciousness - you can map and count the interaction in a brain while thinking (Koch and Tononi use a measure they dub Phi). Because the system has internal causality, Koch argues that it has intrinsic properties (while acknowledging this is not *exactly* a physicists definition of intrinsic, which would require those properties to exist without external stimulus).
Koch starts, as he must, with a definition of consciousness. One which is, by necessity, somewhat vague. Consciousness is experience. Koch does not really elaborate here, beyond a lot of ranting about Dennett being irritating, which is a shame, because clarifying what Koch *doesn't* include in his definition - the sense of internal narrative, and perhaps more crucially, a person's sense of self - is important for determining what causes it, and how confident we are as to whether any given entity 'is conscious'. Koch's fury at Dennis is largely because Dennett rationalises consciousness out of existence, regarding it as an 'illusion' which Koch points out denies the actual experiences of real people. While I am always in favour of having a go at Dennett - who is unapologetic is assuming philosophy will lead understanding and science will follow, which inevitably leads to the conclusion that we know nothing at all - it is also important to note that we all have very specific experiences of being conscious, and there are significant challenges in establishing how much of that specificity is shared. Consciousness in this book is treated as a binary: you have it or you don't have it. There is no discussion of how the differing structure of brains between consciousness might be a spectrum, and hence a wildly differing phenomenon. While there is some discussion about how particular areas of the brain seem involved in both phi and conscious thought - for example, the posterior cingulate cortex - there is no examination of what this might mean for species with different brain structures for experience. At one point, an equivalence between consciousness and affective empathy is thrown out there as if all species which have the first, must have the second (has the author met cats?). There was little exploration of time, despite some intriguing point at the start that hallucinogens are pretty strong evidence that a sense of time is not intrinsic to experience.
Koch presents a compelling argument that Phi is measuring *something*, but not a compelling argument that that something is experience. There is a whole chapter about the relationship between intelligence and consciousness, in which Koch shows indications that Phi correlates with species-level intelligence measures, and some interesting correlations between Phi and patients in various states of conscious impairment - which could also translate to intelligence as well as 'experience'. This is a big issue for the thesis. Phi is demonstrated as something widespread among animals, but - crucially - without being able to link this to a sense of experience, it is feels off.
Throughout the book, Koch treats consciousness as if it is a system designed to interpret reality as exactly as possible. He fails to discuss the various pieces of evidence that suggest that what we are conscious of is not only a limited section of what we are aware of, but that in some cases, it may contradict it (e.g tests that indicate that when we 'decide' to move an arm, we've already started moving it, or brain injuries in which people rationalise their flawed perceptions with events that never happened). Koch argues that we might have "multiple consciousnesses" to explain how we perform actions without being conscious of them, or how split brains work, without really explaining why the simpler solution - that conscious experiences is just one part of brain function not the (lower case) whole - is not preferred. It is hard not to think that this is because Koch wants consciousness to be a fundamental law of the universe, something fundamental, and well, sacred.
In all the book, Koch displays this desire to equate Phi to a property like gravity. One of the sillier parts of the book is when he argues that expecting computer intelligence to simulate consciousness is like expecting a Black Hole simulation to suck the science team running it into the computer. None of these analogies (note my earlier comments about the switcheroo around intrinsic) worked for me, nor did I see the point of them. Koch himself in the past has pointed out how from demonstrable/observable M-theory is. Sometimes this seems to come at the expense of examining the evolutionary purpose of conscious thought. I remain most fond of the attention theory here - that conscious thought enables us to manage considerable attentional brain resource on things of importance. Koch dismisses this in passing - actually one thing that kinda delighted me is that little of what I thought was missing was not acknowledged by Koch, it's just that his responses were unconvincing or simply "yeah, I know, going to say it anyway" in tone.
So it might seem that all I had was complaints. In the end, I was not convinced that the evidence for Phi equalling consciousness was that strong - but neither was I convinced it couldn't. I had a great time reading the book, highlighted the living hell out of it, argued about it, and have left. So. Much out of this review because it is already ridonkulous. I would love to have a rich vegan meal, including wine, with Koch and argue it all through, and if he publishes another book, I'll have it on preorder so fast your head will spin.

[Owlseyes]

http://m.nautil.us/issue/76/language/...

[Arno Klein · Rating 4 out of 5]

I read Christof Koch’s book twice while bedridden with pneumonia, which attests to its clarity and engaging qualities. The author, in writing and interviews, gives the clearest exposition of integrated information theory I have come across.

I appreciate his simple definition of consciousness as experience (p1), stripped of other redundant modifiers like “subjective” or “phenomenological”. I felt, however, that some arguments did not stand on their own. Even for issues that I am in whole-hearted agreement, such as animal consciousness, I felt the author’s reasoning was at times incomplete or unsubstantiated, which is perhaps to be expected in a short, accessible book on such a large topic.

[Linards Jansons · Rating 4 out of 5]

*** 4.5 stars ***

Consciousness.

As long as we’ve got it and it's working OK, we don’t tend to think too much about it. All the same, consciousness strikes me as one of the stranger mysteries of existence, given its surprise appearance in an apparently unconscious universe.

Because, as Christoff Koch himself writes, how does the mental get squeezed out of the physical? Despite all the astonishing advances in neuroscience, we still have no idea how to answer this question (often referred to as "the hard problem"). We can pinpoint with increasing accuracy which areas of the brain correlate with various cognitive and motor functions (NCCs, or neural correlates of consciousness). We can artificially stimulate feelings by poking around in the brain (or taking drugs). But where in all this is the experience? Where does it hide? From the outside we see neurons, synapses and associated bits and bobs sparking away in amazing detail. But how does this observable brain activity translate into the very feelings, sensations and thoughts making up our lived experience? A water into wine and stones into bread miracle that we carry around with us every day.

Christof Koch, who previously worked with Francis Crick (of DNA fame), addresses this miracle through the theoretical framework of Integrated Information Theory (IIT), developed by Italian neuroscientist Guilio Tononi. Koch does not trivialise the mysterious aspect of consciousness, but neither does he regard it as outside the scope of scientific inquiry. In The Feeling of Life Itself, Koch makes a valiant (but not entirely successful) attempt to explain IIT for the general reader. More accessibly he discusses some of the scientific and ethical implications of IIT. I have to confess that this book was not always easy going. All the same, I’ll try to provide a brief overview of IIT before touching on some of the more interesting scientific implications and ethical considerations.

Here’s a helpful starting point. Most theories of consciousness begin with the physical aspects of the brain and then ask how and where consciousness emerges. IIT reverses this, beginning with a phenomenological description of experience, and then going on to ask how physical matter needs to be organised to support these mental aspects. In other words, rather than squeezing brains and asking how the juice of consciousness comes out, IIT analyses the juice itself and asks what a kind of a mechanism could produce such juice. The five properties of consciousness Koch identifies probably don’t mean much without extensive description, but here they are: consciousness exists for itself, it is structured, it is informative, it is integrated and it is definite. The first, for example, states that consciousness is entirely subjective and interior. The last states that the things of experience are completely in, or completely out. There is no half-experience. Each of these 5 postulates of experience, Koch claims, has a physical manifestation which enables it. This is demonstrated through lots of circuit diagrams which requires a knowledge of logic gates, the basis of computer technology, something I didn’t even know was a thing. But now I do.

The upshot is the following: “According to IIT, consciousness is determined by the causal properties of any physical system acting upon itself. That is, consciousness is a fundamental property of any mechanism that has cause-effect power upon itself. Intrinsic causal power is the extent to which the current state of, say, an electronic circuit or a neural network, causally constrains its past and future states”. I sort of get this, but only sort of. The point is that any physical structure supporting consciousness has a kind of power over itself, drawing on past states and determining future states.

Another key aspect is expressed by another of Koch’s compact summaries: “Any experience is identical to the maximally irreducible cause-effect structure associated with the system in that state”. By “maximally irreducible” Koch means that a conscious system is irreducible, more than the sum of its parts, and not comprising the part of a greater Whole (the word “Whole” means any “physical substrate” understood to be conscious). That is, any unified mind cannot be made up of lots of smaller mini-minds. Ten dummies closely huddled together don’t make one brainy guy.

Finally, IIT posits that a conscious Whole consists of a maximum of “integrated information” (hence the name of the theory). This maximal integration is given a measure, named “phi”. IIT predicts that systems with sufficiently high phi are integrated enough to attain consciousness. Details aside, the take away is that it’s not how big the brain is, it’s not how many neurons there are, it’s not even how powerful some processing functions are. It’s how integrated these components are, so much so that a Whole does not permit any subsets (smaller parts of itself) or supersets (larger entities of which it is a part) with more integrated information. And in the brain, far more than the most powerful supercomputer, the level of integration is almost unfathomable. 100 trillion synaptic connections amongst over 100 different kinds of neurons, all in complex relationship with each other.

That’s my feeble summary of the hard parts of this book. Thankfully, the further examples and implications provided by Koch sheds some light on this very dense but fascinating topic. I’m just going to randomly list some of the take-aways that grabbed me.

(1) Not every part of the brain is directly connected with conscious experience. There are entire areas of the brain with no direct responsibility for any specific experiences. The cerebellum, for example, containing a whopping share of the brain’s neurons, is important for muscle control and movement, and for some cognitive functions. But remove it, and consciousness remains. According to IIT this is because despite the sheer quantity of neurons in the cerebellum, they are not wired with a sufficient level of integration. Consciousness, however, seems to be the special function of the cortex, particular the rear section, the so-called “posterior hot zone”.

(2) We already know from Dissociative Identity Disorder that more than one conscious 'entity' can be produced by a single brain. Split brain patients are another example. When Roger Sperry famously severed the corpus callosum to relieve the effects of severe epilepsy, later experiments demonstrated that each hemisphere was effectively independent of the other and could truly be called a 'mind of its own'. According to IIT, therefore, each hemisphere in these cases is not half a consciousness, but a full, albeit different, conscious entity. Each has attained phi max, the necessary level of integration for subjective experience to take place. This leads one to think of the brain like a conference centre, able to host more than one event at the same time, although it also appears that only one event gets 'televised' at a time. That is (to change the metaphor), only one conscious identity is behind the steering wheel at a time while the others are grumbling or sulking or sleeping in the back seat.

Where it gets bizarre is a follow up thought experiment proposed by Koch. Imagine a theoretically (but not practically) possible procedure known as "brain-bridging". If the severed corpus callosum was able to be reconnected, the individual ‘personalities’ of each hemisphere would ‘pouf’ out of existence to be replaced by its former integrated consciousness. But why stop there? If brain-bridging could connect your cortex and my cortex, then at some point during this procedure, our own distinct experiences, memories, and sense of identity would completely vanish, to be replaced by a new, single consciousness incorporating the memories and secrets, knowledge and experience, fears and dreams of both of us, combined as one unified self. Yes, science fiction indeed, but with some skerrick of plausibility.

(3) Koch roasts the brain-as-computer and mind-as-software mythos gripping popular imagination and no doubt promoted by Silicon Valley. But distaste of this dominant metaphor aside, IIT itself demonstrates why computers cannot achieve consciousness. No matter how many circuits are piled up, computers have nothing near the complexity and integration observed in organic brains. One of the most instructive diagrams is a consciousness-intelligence plane. On the intelligence axis computers will, and already do, outsmart humans in many areas. But on the consciousness axis computers remain at zero. Even a fly has more of an inner life, miniscule and annoying as it is.

(4) Ethical considerations crop up throughout the book. In a chapter discussing tools to measure consciousness the plight of brain damaged and disordered patients is addressed. The original electroencephalograph (EEG) designed in the 1940’s (its inventor was testing for telepathy!) has advanced to more recent technology measuring perturbational complexity, the way an impulse travels throughout the brain like a resounding bell. This technique – also developed by Tononi – has shown increasing accuracy in identifying consciousness in the absence of more obvious motor responses. A somewhat disturbing finding is that a small number of patients with zero motor responses are nevertheless experiencing some form of conscious awareness.

(5) The final chapter deals with panpsychism. Koch’s answer to the question “Is consciousness everywhere?” is along the lines of “many but not all things are enminded”. Acknowledging the beauty of panpsychism, and the many famous minds who have endorsed it, IIT suggests that many systems are not integrated enough to support IIT. On the other hand, Koch writes “I now know that I live in a universe in which the inner light of experience is far more widespread than assumed within the standard Western canon”.

This particularly applies to the life of animals and their capacity to suffer, a suffering exacerbated by ruthlessly efficient farming and harvesting techniques. Koch even believes that our treatment of aquatic species (think of boiling lobsters alive) needs rethinking. Koch’s ethical stance promotes vegetarianism as an acceptable minimal option and veganism as a stronger stance. However, he also notes that species do not suffer to the same degree or intensity. “The moral privileges we accord to species should reflect this reality: They are not all situated on the same moral rung” (sorry spiders, bad luck ants). Ranking is necessary as “we must take the graded nature of the capacity to experience into account if we are to balance the interests of all creatures against each other”. The mathematical rigour of IIT (perhaps) provides a way of doing this. On the final page Koch commends the work of Peter Singer in this regard and references Matthew 25 (“what you did for the least of these”) in relation to our non-human sentient relatives.

(6) Koch’s book could give the impression that it explains the question posed at the beginning: how the mental arises from the physical. Yet it appears that IIT does no such thing. There is, in the final analysis, probably no answer why a maximum of integrated information should feel like anything. IIT probably goes further in correlating the physical with the mental, but offers no advances on why this should be so. Koch himself regards such speculations as fundamentally absurd, a case of “trying to peek behind the curtains that hide the origin of creation only to find an endless set of further curtains. I will happily go to my grave knowing that in this universe, IIT characterizes the relationship between experiences and their physical substrate”.

(7) One final observation: If one divided the landscape of consciousness theory into the lands of materialism, dualism, idealism, and panpsychism, then IIT would appear to straddle materialism and panpsychism. Reading Koch's book has furthered my appreciation for how our inner, subjective lives are so utterly dependent on the organic, chemical and physical "substrate" of our brain and nervous system, without detracting for a moment from the enigma of consciousness.

[Shashwat · Rating 4 out of 5]

“Experience is found in unexpected places, in animals both large and small, but not in computers though they speak in tongues.”

I may have overreached with this book. Though Christof Koch brilliantly tries to simplify the complex mind-body problem, the problem itself requires one to be well-versed in multiple fields. Anyway, let me try my best to review what I managed to understand.

Consciousness is, simply, the feeling of life itself. According to the Integrated Information Theory (IIT), which is the leading theory on Consciousness today, Consciousness is definite, informative, integrated, structured, and (most importantly) exists for itself – or has causal effect on itself.
The theory goes on to describe how the neurology of consciousness has allowed scientists to develop consciousness meters based on EEG principles that work well in clinical settings. Most animals, perhaps all, experience life in some form or another, and are therefore conscious. However, he also argues that computers and AI don’t have the same mechanism for consciousness and may never have. Even a perfect software model of the brain may never be conscious, because consciousness is not a type of computation, or a clever hack. It is being itself.

IIT posits that any one conscious experience is IDENTICAL to a maximally irreducible cause-effect structure in the brain and is determined by the causal properties of any physical system acting upon itself.

Multiple feedback loops are essential for an inner experience. Computers and AI work on feedforward circuitry which, though may have many processing layers, is reducible to its components. They do not have sustained feedback to have causal powers over themselves. Conventional digital computers, built out of circuit components (dry hardware) with sparse connectivity and little overlap among their inputs and their outputs do not constitute a Whole. Yes, it may be possible to build computing machinery that closely mimics neural architecture, but we are still far from it.

[Tom Walsh · Rating 3 out of 5]

I wanted very much to find this book valuable having just finished Alan Lightman’ Transcendent Brain in which he speaks highly of Koch’s Theory. And I did admire the essential discussion of his notion of his IIT of Consciousness being pervasive in all beings.

But I wanted him to write more of the Brain’s function in evoking Consciousness than he did, he concentrated too much on the Nuts and Bolts of Neuronal activity and not, at least for me, of his overall vision of Consciousness arising from that low-level activity.

For this reason I found the book unsatisfying though I like what I understand of his Theory. I have to give it only Three Stars but it was an enjoyable read. ***

[Дмитрий Филоненко · Rating 3 out of 5]

In short, author means that consciousness (or feeling of life, or experience) obeys 5 axioms:
any experience exists for itself, is structured, is the specific way it is, is one, and is definite.

The first one means that someone's experience exists for itself, it doesn't need any external observer.
The second one (experience is structured) says about spatial relationships, like on a drawing when you analyze it.
The third one says that experience is informative: contain great deal of details bound together in specific ways.
The fourth one says that experience is integrated, i.e. can't be split on separate independent sub-experiences.
The last one is the most vague, I must admit: "There is a distinct content of consciousness that is “in” while everything else is out, not experienced."

Then Koch claims: "By construction, these five properties fully delimit any experience. Nothing else is left out. What people mean by subjective feelings is precisely described by these five axioms. Any additional “feeling” axiom is superfluous. Is there a mathematically unassailable proof that satisfying those five axioms is equivalent to feeling like something? Not to my knowledge. But I’m a scientist, concerned with the universe I find myself in, and not with logical necessity. And in this universe, so I argue in this book, any system that obeys these five axioms is conscious."

From these axioms the author makes step to the Integrated Information Theory created by Giulio Tononi. In this theory every axiom from above is converted to a causal postulate to which every conscious system has to obey.

The fist axiom will become a postulate which states: to exist intrinsically, any set of physical elements must specify a set of “differences that make a difference” to the set itself. "...for a system to exist for itself, it must have causal power over itself. That is, its current state must be influenced by its past and it must be able to influence its future."

The second axiom has a counterpart postulate which requires from a conscious system to have mechanisms each of which should be able to make a difference in the system, i.e. to change its state. If we present the system as a network of nodes with certain abilities (like in circuits of logical elements) then such mechanisms may be presented by single nodes or pairs of nodes, or combinations of higher number of nodes. These mechanisms in other words have causal power within the system.

The third axiom (information) converts to a postulate which requires from every mechanism to represent a state which is a result of action of other mechanisms (to have a cause in the past) and to have an impact on other mechanisms (to have an effect in future). The higher combined past-future effect a mechanism has, the higher cause-effect power it possesses.

The fourth axiom/postulate (integration) requires from cause-effect structure to be irreducible. "The system can’t be reduced to independent, noninteracting components without losing something essential."

And finally the last axiom is represented by the Exclusion postulate. This one claims that only the circuit that is maximally irreducible exists for itself, rather than any of its supersets or subsets. All overlapping circuits with smaller values of integrated information are excluded.

With all that said, Koch asserts that ANY system which satisfies these postulates will be conscious.

A significant part of the book is devoted also to other related topics, starting from difficulties to define if a certain patient with heavy brain injuries is still conscious or not and finishing with explanation why currently booming AI software is not conscious and will not be able to show any signs of consciousness unless it's architecture will be changed from feedforward (which have zero integrated information according to IIT) to something more sophisticated.

Well, I'm not a scientist. And this book is obviously written for laypeople like me, i.e. without heavy rigorous proofs, strict chains of conclusions, etc. So my scepticism might be caused by this. But anyway.
Axiom is a very strong notion in science. It's a thing which should be absolutely obvious and thus very simple. If we take axioms known to everyone from school - those of Euclidean geometry - while they seem extremely obvious and straightforward, even they are controversial (well, one of them at least, about two points and a straight line). Can I agree that axioms suggested by Koch are obvious? I can't.

Yes, they seem reasonable and logical. But that's the problem. Can we use logic here that easily? Is it the same - our experience and our expression of experience: in language first of all, or in any other conceivable signs? Second axiom, about structure in experience, requires that we have a kind of coherent picture of what we perceive with all spatial relationships, etc. But our expression of experience is pushed forward through a few processing blocks: at least analytical one (logical analysis of what I have seen or heard, etc) and speech 'generator' (when ready analyzed information is converted to another representation made by, again, strict and logical rules). My description of a room experienced by myself now will of course be rather accurate, with all due perspective layers (background, foreground...), what is to the left, what is to the right, and so on. Is it really how I experience it or is it a result of post-processing needed either for reflection over it or for conveying it to someone else?

Then the fourth axiom about integration. Well, maybe it is so. But is it obvious? I'm pretty sure we don't experience life as we see it in a movie: a linear sequence of actions, everything is chained. Even in literature a flow of consciousness invented (?) by Faulkner and Joyce, seems closer to how it really is. Can we call this integrated experience? Or is it actually a series of independent sub-experiences?

The move from axioms to postulates is not convincing either. "For a system to exist for itself, it must have causal power over itself". Well apparently our brain has these traits: very sophisticated interconnected network. But first, really for itself? Well, my experience is my experience, it doesn't need external observer, right. But what if external observer is within me? Just another part of the brain which uses results of conscious analysis? Second, can't be such a wildly interconnected network of neurons a byproduct of a blind evolution? Is it really necessary for consciousness to have such a sophisticated substrate? I don't see that consciousness should infer 'by axiom' a network with causal power over itself. And moreover, I don't see why a network with such properties must be assumed conscious.

Connection between second axiom and postulate is absolutely vague. Structure in experience on one side and requirement for mechanisms in a system which should be able to make a difference in the system, on the other side, do not seem directly related.

The book mentions a criticism to the IIT, in particular an objection presented by Scott Aaronson. In it he shows an example of a regular grid of XOR logic gates which according to IIT has very high integrated information and as a conclusion must experience something, to have a consciousness. This looks weird. Koch with delight presents the answer of Giulio Tononi to this criticism. In short, this answer suggests to consider a blanc wall, which is featureless only to an external observer but inside is full of relations... Well. The author means that this answer was brilliant and strengthened the theory even more. I was curious and followed a link to this discussion in a blog of Aaronson. He presented four arguments for why such an answer is far from being convincing. The same interesting was to read comments of other interested in this question.

Quite a controversial book. Though with interesting ideas and highlights on the current state of neuroscience.

[Vikram Chalana · Rating 4 out of 5]

An interesting book which tries to give a formal definition of consciousness. The concept laid out is something called integrated information theory (Iit). A closed system that exists for its sake is defined as conscious. The more complex the system, the higher the degree of consciousness. Thus, human brain has high consciousness index, while a mouse brain has a lower one. An atom may also be conscious, but with a with a very low index. Artificial Intelligence has a very small degree of consciousness because the system does not exist for itself. So, consciousness exists everywhere, but according to Koch consciousness emerges from matter.

The end of the book was quite a bit less profound than the beginning. The actionable insight on this work on consciousness seems to be that we should all be vegetarians since every animal we eat is potentially conscious and can feel pain.

[mono]

α: mitpress
Σₐ: IIT(integrated info theory) + panpsychism = not computable consciousness

ᴥ: Wolfram on consciousness

[Nikki · Rating 4 out of 5]

I read this entire book, but mostly out of order. It was fascinating, but in order to fully understand it, I had to start with the introduction and first chapter. Then I skipped to the last chapter, and when it got to the synopsis, would go back and read the chapter about each idea. I think I needed more context to grasp some of the information.

[Steven H · Rating 5 out of 5]

THE NEUROSCIENTIST EXTENDS HIS IDEAS EVEN FURTHER

Christof Koch is a German-American neuroscientist who is president and chief scientist of the Allen Institute for Brain Science, and was formerly a professor at the California Institute of Technology.

He wrote in the Preface to this 2019 book, “How does the mental relate to the physical? Most assume that the mental emerged from the physical when the physical became sufficiently complex… Yet are we really to believe that until that point in time… the world ‘[was]… quite properly speaking not existing?’ Alternatively, perhaps the mental was always present, allied with the physical, but not in a form readily recognizable? Perhaps consciousness predates the arrivals of big brains? This is the less-travelled road that I will take here… Then there is the urgent question of the day---can computers experience anything?... human-level artificial intelligence may come into the world within the lifetime of many readers. Will these AIs have human-level consciousness to match their human-level intelligence? In this book, I will show how these questions… are now being addressed by scientists… the science of consciousness has seen dramatic progress over the past decade… Much cognition occurs outside the limelight of consciousness. Science is bringing light to these dark passages where strange, forgotten things live in the shadows.” (Pg. xi-xii)

He continues, ‘After considering why consciousness evolved, [this book] turns to computers. The basic tenet of today’s dominant faith … is that digital, programmable computers can, in the fulness of time, simulate anything, including human-level intelligence and consciousness… According to integrated information theory, nothing could be further from the truth. Experience does not arise out of computation… While appropriately programmed algorithms can recognize images, play Go… and drive a car, they will never be conscious… because it lacks the intrinsic causal powers of the brain… It will claim to have experiences, but that will be make-believe---fake consciousness… Consciousness belongs to the natural realm… it has causal powers… I will show that the intrinsic causal powers of contemporary computers I puny compared to those of brains.” (Pg. xiii-xiv)

He states in the first chapter, “Consciousness is experience… any experience, from the most mundane to the most exalted… I use these two words [awareness and consciousness] interchangeably. I also do not distinguish between FEELING and EXPERIENCE… As I use it, any feeling is an experience. Collectively taken, then, consciousness is lived reality. It is the feeling of life itself… Without experience, I would be a zombie, a nothing to myself.” (Pg. 1) He adds, “What would be simpler? Consciousness is the way the world appears and feels to me.” (Pg. 3)

He outlines, “every conscious experience has five distinct and undeniable properties: each one exists for itself, is structured, informative, integrated and definite. These are the five essential hallmarks of any and all conscious experiences, from the commonplace to the exalted, from the painful to the orgiastic.” (Pg. 9)

He notes, “over the past century the scientific perspective on consciousness has undergone a curious inversion: consciousness has been evicted … and has migrated downward. There is nothing refined, reflective, or abstract about … [a] vast multitude of experiences… If this is the true state of affairs, then it is overwhelmingly likely that not only humans but many and perhaps most animals, small and large, experience the world. Indeed, it turns out that our most refined cognitive abilities, such as thinking or being creative, are not even directly accessible to experience.” (Pg. 34) He continues, “Creativity and insight are two key aspects of intelligence. If these are inaccessible to conscious introspection, then the relationship between intelligence and consciousness is not straightforward. Maybe these are really two different aspects of the mind? Isn’t intelligence ultimately about acting smart in the world and surviving, whereas experience is about feeling? Under that view, intelligence if all about doing, while experience is about being.” (Pg. 36)

He states, “I see no need to invoke exotic physics to understand consciousness. It is likely that a knowledge of biochemistry and classical electrodynamics is sufficient to understand how electrical activity across vast coalitions of neocortical neurons constitute any one experience. As a scientist, I keep an open mind. Any mechanism not violating physics might be exploited by natural selection.” (Pg. 69)

He considers the philosopher David Chalmers: “Chalmers asks whether the existence of zombies is at odds with any natural law… Chalmers concludes that no natural law precludes the existence of such zombies… conscious experience is an additional fact above and beyond contemporary science. Something else is needed to explain experience. He acknowledges the existence of … empirical observations that link the material world to the phenomenal one… But WHY certain bits and pieces of matter should have this close relationship to experience is a mystery that is hard, and perhaps even impossibly hard, to solve. Integrated information theory (ITT) … doesn’t try to squeeze the juice of consciousness out of the brain. Rather, it starts with experience and asks how matter must be organized to support the mental.” (Pg. 73-74)

He explains, “You have arrived at the heart of the book… According to integrated information theory (ITT), consciousness is determined by the causal properties of any physical system acting upon itself. That is, consciousness is a fundamental property of any mechanism that has cause-effect power upon itself… The theory takes the five phenomenological axioms of experience that I introduced in the first chapter … and formulates for each one and associated causal postulate, a requirement that any conscious system has to obey… these causal powers are identic to conscious experience, with every aspect of any possible conscious experience mapping one-to-one onto aspects of this causal structure.” (Pg. 79)

He acknowledges, “we still do not know what survival value is attached to experience. Why are we not zombies, doing everything we do but without any inner life? On the face of it, nothing in the laws of physics would be violated if we didn’t see, hear, love or hate but still acted as we did. But here we are, experiencing the pains and pleasures of life.” (Pg. 119-120)

He observes, “Some … argue that consciousness has no causal role at all. They … argue that feelings have no function… I find this line of argument implausible. Just because consciousness isn’t needed to accomplish a well-rehearsed and simplistic laboratory task in no way implies that consciousness has no function in real life… Consciousness is filled with highly structured percepts and memories of sometimes unbearable intensity. How could evolution have favored such a tight and consistent link between neural activity and consciousness if the feeling part of this partnership had no consequences for the survival of the organism?... If experience had no function, it would not have survived this ruthless vetting process.” (Pg. 120-121)

Returning to the issue of computers, he says, “Fast forward a few decades into the future when … anatomically accurate whole-human-brain emulation technology … can run in real time on computers. Such a simulation will mimic the synaptic and neuronal events that occur when somebody sees a face or hears a voice. Its simulated behavior … will be indistinguishable from those of a human. But as long as the computer… resembles in its architecture the von Neumann machine … it won’t see an image… it won’t experience anything. It is nothing but clever programming. Fake consciousness---pretending by imitating people at the biophysical level.” (Pg. 150)

He admits, “To the extent that I’m discussing the mental with respect to single-cell organisms let alone atoms, I have entered the realm of pure speculation, something I have been trained all my life as a scientist to avoid. Yet three considerations prompt me to cast caution to the wind. First, these ideas are straightforward extensions of ITT… to vastly different aspects of physical reality… Second, I admire the elegance and beauty of this prediction. The mental does not appear abruptly out of the physical… Third, ITT’s prediction that the mental is much more widespread than traditionally assumed resonates with an ancient school of thought: PANPSYCHISM.” (Pg. 160)

But then he adds, “But panpsychism’s beauty is barren. Besides claiming that everything has both intrinsic and extrinsic aspects, it has nothing constructive to say about the relationship between the two. Where is the experiential difference between one lone atom zipping around in interstellar space, the hundred trillion trillion making up a human brain, and the uncountable atoms making up a sandy beach? Panpsychism is silent on such questions… Most importantly, though, ITT is a scientific theory, unlike panpsychism.” (Pg. 162) He continues “Finally, panpsychism has nothing intelligible to say about consciousness in machines… [Whereas] ITT offers a principles, coherent, testable and elegant account of the relationship between these two seemingly disparate domains of existence… grounded in extrinsic and intrinsic causal powers.” (Pg. 166)

He concludes, “ITT can rank species according to the quantity of their consciousness, a modern version of the ‘Great Chain of Being.’ … I understand the squeamishness that such a ranking provides. However, we must take the graded nature of the capacity to experience into account if we are to balance the interests of all creatures against each other… We should treat all animals as being conscious, as feeling what-it-is-like-to-be… One day, humanity may well be judged for how we treated out relatives on the tree of life.” (Pg. 172-173)

This book will be of keen interest to those studying consciousness, and neuroscience.

[P K · Rating 1 out of 5]

Let me preface by saying I don’t typically find books about consciousness interesting because they tend to be self-indulgent exercises in circular babble, and those are the scientific ones (leaving aside completely books not rooted in the scientific method). Additionally, they tend to present few if any actionable pieces of information and do a poor job rooting the findings in practical significance, which for me is a must. To me, the field of consciousness is just another life science, albeit one that’s notoriously hard to investigate, and I hold this field to the same standards as when I read other neuroscience books.
That said, this book avoided many of those pitfalls. Koch writes at length about the practical implications including finding out how much animals can suffer, which comatose patients are likely to wake up again, and excitingly, whether it’s likely that AI systems will have consciousness. That’s all quite cool. My other favorite aspect of the book was the description of the “zap and zip” method. The idea of a single numerical metric of conscious ability is really cool and obviously has a lot of useful applications. However, I was familiar with this material prior to reading this book, and it is not Koch’s invention. The potentially novel parts of this book were so choppily and cursorily explained that the book as a whole was a total slog. Koch presents the five axioms of IIT with absolutely no explanation of why we should accept these premises. Moreover, after carefully reading the relevant sections two or three times in a row, it was not clear to me how one axiom in particular was non-overlapping with another particular one. The section about the neural correlates of consciousness also seemed really arbitrary and poorly explained to me. For instance, it wasn’t clear to me why a change in personality constituted a non-change in consciousness but a change in remembering what it felt like to see color did, if we accept as Koch wants us to that consciousness is just what it feels like to experience something.
Overall, I don’t feel like I learned anything at all from this book, but it did take me a really long time to read and sort of understand.

[Peter Gelfan · Rating 4 out of 5]

Koch tackles the “hard problem” of philosophy and neuroscience: consciousness. While some dispute that such a thing exists separate from or in addition to the neurological functions that enable us to deal with our bodies and environment, Koch gives consciousness a specific definition: our inner experience of life. A thermostat may in a sense be aware of the ambient temperature and thereby control heating and cooling devices, but does it experience anything? You know that you experience things, but you cannot convey your experience, along with all its feelings and nuance, to another. Words or other forms of expression can impart the gist of your experience not precisely recreate it. In fact, none of us can know with certainty that anyone else experiences anything. They can say they do, but your cell phone can say that too. Which puts neuroscience and philosophy in an embarrassing position: they cannot examine firsthand the essence of what it means to be conscious. We can share facts, images, sounds, ideas, and so on but I can never experience what it feels like to be you.

Dualists will say here’s where spirit, soul, comes in to make life complete. Physicalists say either that consciousness doesn’t really exist or that it’s an emergent property of brains and possibly other complex systems. Panpsychists believe that consciousness is an integral part of all that exists. Koch has a different explanation. Integrated information theory, IIT, which he helped develop, is his answer. It begins by breaking experience down to its essential components and then studying how they work (or not) in brains (and computers).

The book is fascinating. In places it’s also quite technical. For some of the diagrams, I confess I lifted the hood, stared down inside, then closed the hood again. But overall, the theory makes sense, is testable, and doesn’t require a leap of faith to any ism. He writes well with a straightforward manner.

[Florent Diverchy · Rating 3 out of 5]

This is the kind of scientific book that generate a lot of frustration in me. Not so because of the content, but rather because of the way it is written:

* The author regularly declares, in the main text or in the footnotes, that it's obvious that his theory is right and that the other theories are wrong, without any demonstration.

* The most cited author throughout the book is ... himself, and it feels like "Look, I'm right because this other guy (which is me, but that's just a detail) is saying the same thing".

* Some statements would have needed a few extra comments:
- Is consciousness really irrefutable? No need to talk about Boltzmann brains?
- The experiences of our ancestors are encoded in our DNA? Care to explain a bit?

More disturbing was the fact that the end of the book seems to contradict the begin, and even the subtitle of the book. I sometime had the impression that the author was not entirely understanding his theory, or that he was drawing the wrong conclusions from him.

The book starts with the postulate that "consciousness = experience".
However, in the second part of the book, it becomes obvious that consciousness emerges due to the fact that our neurons are arranged in networks which are not just feeding forward, but are also feeding backward, in a bidirectional way.
This rather means that "consciousness = remembrance of experiences past".

This also leads to the fact that (as the author himself is saying): we can't compute consciousness with our current feeding forward neural networks, but we could be able to do so with recurrent , feeding backward neural networks.

In that case, why subtitling the book "Why consciousness can't be computed?", while a more correct abstract of the book would be "Why consciousness can't be computed YET?".

Is this a marketing move? A clickbait?

For all these reasons, I can't give more than 3 stars to this book.

[John · Rating 2 out of 5]

I actually quit about halfway through so can't comment on the latter half.

Ultimately, it felt like a rambling argument that didn't establish its positions, like something your friend might say while drinking a beer, in particular, a friend who had a passing acquaintance with many aspects of the topic but not one sufficient to get to the meat of issues.

Was kind of a bad sign early on when he first misrepresented Dennett's position on qualia (in a very one-off way), then called Searle a doyen, using a very cut-up line, where Searle's positions are mainly interesting in that they're intuitionist constructions that show us how intuitionism leads us astray when we break apart his arguments (although this wasn't Searle's intention, who appears to have believed in the strength of his arguments -- note I'm referring specifically to his AI-related arguments).

I quit the book because I figured if the things I knew about had these flaws in argumentation, it put into question any trust I could have in assertions he might later make on anything I didn't already know about. If, on the other hand, I already knew about everything he was going to talk about, why read it? Note, it would have been fine if he had argued well from axioms I might disagree with, as long as he'd been clear on that, but the book lacked that precision.

IMO, the book treats its subject too breezily. One could say it's a popular treatment, but I'm not sure what one is getting out of it then. Had it been given more focus, it could have been a very different thing. Koch seems to have the credentials after all.

[Evgeniya Kondrashina · Rating 5 out of 5]

The first book on the subject of consciousness I have read so far that proves in a rigorous and scientific way that consciousness of varying degree is a property of many connected systems, including most living beings. It argues for a sort of scientific, mathematical panpsychism.

1. The author is a mathematically trained neuroscientist who understands all the ins and outs of the computational approach to modelling the brain's workings. 2. He explains the reductionist nature of such an approach and demonstrates that mathematics is insufficient to explain the nature of consciousness 3. The way current computers and AI are being programmed is fundamentally different from how the brain operates therefore computers will not be conscious 4. He provides a detailed discussion of Integrated Information Theory (ITT) - the best theory available so far to account for the phenomenon of consciousness that allows to compare levels of consciousness between living creatures 5. Consciousness is different to intelligence

Chapters are short and to the point, logically structured and very well written. For those seriously interested in the latest scientific thinking on the problem of consciousness, this book is must read!

[Deniz · Rating 1 out of 5]

This book is pseudoscience. It starts with a chapter titled "What is consciousness ?".

And the definition is: "Consciousness is the way the world appears and feels to me." Ooh, I feel enlightened.

So who can have consciousness according to Koch except humans ?
Insects are conscious.
A lump of nerve cells are conscious.
Plants are conscious.
Bacteria ? Surely.
Atoms ? A tiny bit.

According to Koch everything is conscious. EXCEPT COMPUTERS !!! They can never be conscious. Not now. Not in the future. Never. So we can have a super intelligent AI with emotions that can converse about philosophy. But we can destroy or torture it without problem because IIT says it can't be conscious.

So taking an antibiotic is worse than torturing hosts in Westworld it seems. Here's your conclusion for the book. It should have been named "DUUUDE WHAT IF EVERYTHING IS CONSCIOUS (except computers)"

For a proper critique you can read Scott Aaronson's post: Giulio Tononi and Me: A Phi-nal Exchange

[Paul · Rating 4 out of 5]

I heard about this book from an interview with Koch on the excellent Brain Science Podcast and was intrigued to understand more about IIT (Integrated Information Theory).

I found the book to be easy to read and with some good attempts to explain the core concepts of IIT, which in turns out is less about the information content of thought and more about the causal structure of the system being the source of consciousness.

I quite liked how Koch progressed from some key features of subjective experience, to the core theory, to implications in terms of consciousness in humans, other animals, computers and the universe generally. He thinks computers are a long way off having enough connectivity or subtlety of signalling to have any consciousness, though this is not to be ruled out if they get better at replicating neuronal function.

[Gary Moreau · Rating 4 out of 5]

In this book Koch attempts to explain/explore consciousness through the use of neuroscience, advanced mathematics, the highest levels of reason, and biology. In total it amounts to what is known as Integrated Information Theory and he attempts to use it to shed light on the “mind-body problem” that all stripes of thinkers have long struggled with.

“Consciousness is experience,” a perfectly acceptable explanation to my way of thinking. And experiencing something is clearly different than simply doing something. And we can all agree, I think, that living is more than just doing.

Ultimately Koch differentiates between intelligence and consciousness. It’s a sound distinction that helps to put the “AI-human problem” (my own version of the mind-body problem), in its proper perspective. Machines can be very intelligent because they are very fast at performing calculations. But they will never be human. They will never have consciousness, or experience their intelligence, because the way they work is entirely different than the way humans do.

The most powerful computers today are nonetheless still binary and process information in a very linear, hierarchical way. (Quantum computers are not in widespread use yet but primarily represent more speed, not a fundamentally different mode of processing.) The human mind, on the other hand, does not process information in a linear fashion. While its mode of calculation is less vertical or linear, such simplicity allows for much more complicated and nuanced calculations.

The distinction has, I think, very important implications for society at large. We worship computers in ways they don’t deserve and is not healthy to human progress. As computers get smarter (i.e. faster) we are dumbing ourselves down. In business, for example, we are focusing on analytics and Big Data and walking away from intuition, experience, and creativity. Intelligence may create wealth, but it will destroy the human experience if not coupled with consciousness. There is no “mind-body problem.” We need them both to live a rich life.

Which is why I fear that my daughters’ generation will not be as skeptical of AI as it should be. It can and will do some things very well. But it can’t do everything. And if we let it, it will destroy the beauty of life as I have had the privilege to know it. It will destroy joy and the most rewarding forms of learning.

The distinction in how machines and people process information also explains why we have never really discovered where consciousness resides. In computers it is all very simple. The intelligence resides in the processor.

There are several implications of all of this according to Koch. One is that machines will never have consciousness; at least not for a very, very, very long time. The second is that we will never be able to upload ourselves to the cloud or some other artificial form of storage. And the third is that consciousness is probably far more widespread than the human race. While the degree of consciousness may vary widely based on the complexity of the central nervous system, much of all cellular life probably has some level of consciousness, which, of course, has several lifestyle implications, including what we eat and how we treat animals. (I can sign on to that type of thinking.)

Unfortunately, the prose is more tangled than the human neural network. It is, at times, like wading through waist-deep paste. You must be willing to plough through sentences like this: “Causal power is not some airy-fairy ethereal notion but can be precisely evaluated for any physical system, such as binary gates implementing Boolean logic or all-or-none neurons in a neural circuit.”

All told, I believe the book makes some very important and timely observations. I only wish it had done so in a more approachable literary fashion. As it doesn’t I fear that it will be overlooked in the broader community that most needs it.

[Freddie Bennet · Rating 4 out of 5]

How does matter produce mind?

The problem of linking these two utterly distinct worlds has stumped academics since Plato's time. Neuroscientists, physicists and psychologists alike have failed to make any fundamental progress in answering the question, while philosophers have termed the issue of consciousness "the hard problem", illustrating the extent of the mystery.

With the birth of tools like the MRI brain scanner and the EEG, it has become possible to monitor the neural activity which underpins all forms of conscious experience, allowing neuroscientists to track the footprints of consciousness in the brain. Every possible conscious experience, whether it be the perception of a friend's face, the admiration of a sunset, a spike of anger, the surfacing of an old memory or a chain of reasoning, no matter how subtle or explicit, is always underpinned by specific neural mechanisms. The minimum neuronal activity required for any one conscious experience is referred to by neuroscientists as a "neural correlate of consciousness". In recent decades, it has been a major venture of neuroscience to uncover some of these minimal circuits for the myriad variations of conscious experience. Over are the days of the proverbial ghost in the machine as philosophers like Descartes once postulated; we now know that consciousness is inextricably tied up with the brain.

The richness of an organism's conscious experience is determined by the number of neural correlates of consciousness facilitated by that organism's brain. Consciousness does not require the most complex of these mechanisms, such as holding an interior perspective, higher forms of reasoning, broad emotional ranges and metacognition. To be conscious simply means to be capable of having any form of experience. More than just a yes-no binary, consciousness is a spectrum. I used to believe that what was mysterious about consciousness was its higher forms, but without doubt, what is truly mysterious is experience at its most basic and fundamental form. How is it, that a structure of atoms, molecules, proteins and cells suddenly becomes able to feel something, and to have any flavour of experience at all? How is the world of the subjective suddenly birthed by the world of the objective? The true mystery of consciousness lies in the tiny gap between nothingness and something.

An implication is the notion that all animals with brains have forms of conscious experience, proportional to the complexity of the organism's brain. You would be hard pressed to convince me that a dog, for a example, does not have some form of consciousness. We see that dogs (and other animals) have their own distinct personality traits and emotional ranges. Animals show curiosity for life, they experience excitement, affection, happiness. Animals are shown to mourn friends and family, they can sulk and become irritated. The findings of neuroscience increase our accountability for animal rights, learning that even without the most complex forms of consciousness, they still undeniably experience this world, and this experience entitles them to basic ethical rights.

Understanding these minimal correlates of experience will also help us detect consciousness in comatose, vegetative, mute and incapacitated patients, allowing us to identify "stranded minds in damaged brains." The neurological research into consciousness undoubtedly has a number of positive fuctions, and with time we are likely to learn the neural correlates responsible for every facet of consciousness. However, even if we develop a complete understanding of what mechanisms cause which subjective experiences, we will still be at a loss for why. Just why, does a particular coalition of neuron firings result in our subjective experience of the colour red, a powerful surge of emotion, or the feeling of life itself? In line with the past 2,500 years of consideration, abstraction and research, the "hard problem" of why experience is possible will remain.

[Sam · Rating 4 out of 5]

“I think therefore, I am” - Descartes
“If I am mistaken, I exist” - Saint Augustine

Renowned neuroscientist Christof Koch explores the complex, multifaceted nature of consciousness in "The Feeling of Life Itself." He posits that consciousness is fundamental, prior to physics and phenomenology and that subjective experience is the foundation of reality.

The concept of qualia, or the elusive experience, is central to Koch's inquiry. He examines Integrated Information Theory (IIT), which suggests that consciousness arises from integrated information in neural networks. Additionally, Koch discusses Neural Correlates of Consciousness (NCC), the minimal neural mechanisms sufficient for specific conscious percepts.

Interestingly, Koch highlights several striking facts about consciousness and the brain. Did you know that Rene Descartes famously argued that dogs do not feel pain and even performed vivisections on rabbits and dogs? In contrast, modern science reveals remarkable similarities between human and animal consciousness. For instance, human and chimpanzee genomes differ by only 1% yet these variations shape distinct species characteristics and the neurons of mice and humans share similar morphologies.

Furthermore, Koch notes that the brain contains approximately 86 billion neurons, with 69 billion in the cerebellum and 16 billion in the cortex. The remaining structures, including the midbrain and brainstem, account for only about 1% of all neurons. The human brain also boasts an impressive 100 trillion synapses.

The book explores consciousness in animals, highlighting evolutionary continuity across species. Koch challenges human exceptionalism, arguing that consciousness is ubiquitous and supporting panpsychism. He examines altered states of consciousness including physiological states like wakefulness, REM sleep and deep sleep, as well as pathological conditions such as coma and vegetative state. He discusses the challenges of diagnosing vegetative state patients and the complexities of locked-in syndrome and minimally conscious states.

Various theories of consciousness are presented, including quantum gravity theory (Penrose) and IIT, which links ontology, phenomenology, physics, and biology. Giulio Tononi's five phenomenological properties – existence, structure, specificity, unity, and definiteness – are also explored.

However, Koch's treatment of Eastern philosophical traditions, particularly Hinduism, feels somewhat superficial. He fails to capture the nuances of the Sat-Chit-Ananda framework. This ancient Indian philosophical approach posits that consciousness (Chit) is the fundamental reality beyond duality and space-time constraints and is characterized by existence (Sat) and bliss (Ananda). I suggest to those interested to listen to the lectures by Swami Sarvapriyanand on consciousness. Available on YouTube. Also, while Koch provides some context, I found the neurological aspects challenging without prior knowledge of brain function and anatomy.

The book concludes with discussions on all-pervading aspect of consciousness, consciousness in AI and the futuristic neurotechnologies such as brain bridging (linking multiple brains) and mind blending - a reversible connection that raise intriguing questions about the potential for collective consciousness and the boundaries of human and AI experience.

The book offers a profound exploration of consciousness, challenging readers to reexamine their understanding of experience of the world. My review may read more like a summary but I believe this reflects the book's unique nature. As Koch writes that consciousness topic is often absent from textbooks, leaving a significant knowledge gap. In my opinion, "The Feeling of Life Itself" addresses this serving as a comprehensive textbook on consciousness.

[Craig Martin · Rating 4 out of 5]

Christof Koch, who worked for many years with Francis Crick of DNA fame, reckoned that it would be possible to identify the specific parts of the brain that give rise to consciousness in humans. He (and Crick) called these areas the Neural Correlates of Consciousness NCCs). In recent years, while holding onto the search for NCCs, he has become a supporter of a theory of consciousness called the Integrated Information Theory, espoused by Tononi, an Italian superstar. A consequence of Koch and Tononi’s work in IIT is that regardless of how ‘intelligent’ computers and their underlying deep learning models (such as GPT3 and its progeny) become, they will not ever be conscious (at least not using current silicon). Although they are huge, they do not have the depth of interconnections. According to Koch, one of the densest areas of neurons in the human brain, the Cerebellum, containing 80% of the brain’s 86 billion neurons, is not an area of NCC.

The book is an exciting read, and Koch is a witty writer. But the chapters detailing IIT will lead the reader into a mind-swamp. I read them three times and have read several academic papers on the subject, but I still only have a skin-deep understanding.

IIT computes a value - called Phi - for anything. If Phi is zero, the object has no consciousness. Above zero, however, there is an implication for some feelings of self. Koch acknowledges that this could. Be a modern-day form of Panpsychism - everything has consciousness - but points out the logical/mathematical ideas of IIT provide a framework to say why something can or cannot have consciousness. There is even some Python code to allow someone to calculate Phi themselves.

One practical derivative of IIT, described by Koch in the book and elsewhere, is a form of consciousness meter - this allows neurosurgeons to gauge the level of consciousness in coma-like patients. This must be some comfort for those faced with difficult hospital bedside decisions.

The most vital point to emerge from Koch’s book is that AI may be intelligent, at least perhaps someday, but it is not likely to be conscious. He clarifies that Alan Turing’s famous test is not designed to identify intelligence only and not to suggest the feeling of life itself can ever reside in a silicon substrate, regardless of how many deep layers of artificial neurons are therein represented. He gives an intriguing analogy: physicists can safely model the behaviours of a black hole without fear that the machines they are working on will be pulled through the event horizon and crushed to nothingness.

Koch is trying to bring meaning to the world through his lifelong study of consciousness. His humour in writing extends, or is a projection, of his attitude to life. He is prepared to take risks. He was an earlier advocate of the power of Gamma Waves - these 40-100 hz cycles that have previously been thought to be the signals of consciousness. He seems to have abandoned this. His work with Crick gave him the confidence to bet with David Chalmers - the Australian Philosopher - that by June 2023, there would be a clear identification of the ‘neural correlates of consciousness’ - something that Chalmers thought was not possible. He may lose the bet but has also made other long wagers. He is a polyglot of science - as the field of Consciousness study has few boundaries - and through the book, brings across his knowledge of biochemistry, neuroscience, and artificial intelligence, in a thought-provoking way.

[David · Rating 4 out of 5]

***.5

The author is a neuroscientist and strongest when he sticks to his primary field of expertise. The early chapters provide a straightforward definition of consciousness as experience, and uses this to establish a common-sense approach to developing a theory of mind. He then delves into details on how the brain works and how consciousness emerges from the physical substrate and connectivity.

The second part focuses on IIT (Integrated Information Theory) and that's where he lost me a bit. The basic premise that there is a single variable that represents the degree to which a system is self-aware is pretty neat, but he didn't do nearly as thorough a job of justifying the underlying basis for the theory as he had in the first part. He seemed to gloss over some of the more difficult/controversial parts rather than attempt to justify them.

This tendency to hand wave away criticism became increasing problematic as the book progressed. He spent a good deal of time in Part 1 to demonstrate why he thinks animals have consciousness too, but when it comes to denying that potential to computers, it's more an assertion than a substantiated argument. It seems like his aversion to the idea of sentient computers is that it's "icky" to him, with personal opinions supplanting science:

"The mythos that life is nothing but an algorithm limits our spiritual horizon and devalues our perspective on life, experience, and the place of sentience in time’s wide circuit."

He starts the discussion about AI by referencing Gödel and the Church-Turing Thesis, but then immediately drops it without elaboration or drawing any conclusions, and instead pivots immediately to describing a couple of current machine learning applications, then points out the physical differences between brains and silicon-based chips, which is entirely besides the point. There's also a diagram of 2 different neural networks with equivalent inputs and outputs, and supposedly one of them obviously has integrated information ("intrinsic causal powers") and the other doesn't. I read the section three times but was unable to follow the reasoning.

He then makes things worse with the ridiculous analogy that since computer simulations of the weather don't make it rain, brain emulations are therefore unable to achieve consciousness, which he attempts to justify in a footnote with a reference to John Searle's famous Chinese Room argument (which seemed to me entirely irrelevant to the point).

I rounded up the rating because despite my disagreeing with his main conclusion, he then went on to differentiate clearly between the scientific approach of IIT and the mystical elements of panpsychism, before concluding with an exhortation to treat other living things (except AI!) as at least somewhat conscious and hence worthy of compassion and better treatment than they currently receive.