Something It Is Like to Be You

Su propuesta fue buscar los correlatos neurales de la conciencia, es decir, los patrones específicos de actividad cerebral que acompañan a una experiencia consciente concreta.

Their proposal was to search for the neural correlates of consciousness, meaning the specific patterns of brain activity that accompany a particular conscious experience.

No el por qué, sino el dónde y el cómo.

Not the why, but the where and the how.

And around the same time, Benjamin Libet was running experiments that were, honestly, deeply unsettling to a lot of people.

Because what he found suggested that our sense of deciding to do something might come after the brain has already started doing it.

Los experimentos de Libet son icónicos y todavía se debaten.

Libet's experiments are iconic and still debated.

Pidió a los participantes que movieran la muñeca cuando quisieran y que anotaran en qué momento tomaban la decisión consciente de hacerlo.

He asked participants to flex their wrist whenever they felt like it and to note the moment they became consciously aware of deciding to do so.

Descubrió que había una señal eléctrica cerebral, el llamado potencial de preparación, que precedía a la conciencia de la decisión en unos trescientos a quinientos milisegundos.

He found there was a brain electrical signal, the so-called readiness potential, that preceded the conscious awareness of the decision by around three hundred to five hundred milliseconds.

El cerebro, en apariencia, se adelantaba a la mente.

The brain, apparently, was ahead of the mind.

Which, if you take it at face value, is a frontal assault on the idea of free will.

The choice has already been made before you know you're making it.

I remember reading about this and feeling genuinely destabilized for about a week.

La interpretación más extrema, eso sí, es probablemente incorrecta.

The most extreme interpretation is probably wrong, though.

Investigaciones posteriores han cuestionado si el potencial de preparación mide realmente la 'decisión' o simplemente la preparación motora general del cerebro.

Later research has questioned whether the readiness potential actually measures the 'decision' or simply the brain's general motor preparation.

Además, el propio Libet señaló que los participantes podían vetar el movimiento en el último instante, lo que sugería cierto margen de agencia consciente.

Moreover, Libet himself noted that participants could veto the movement at the last instant, suggesting some margin of conscious agency.

Pero el malestar que describes es productivo.

But the discomfort you describe is productive.

Nos fuerza a preguntar qué significa exactamente 'decidir conscientemente'.

It forces us to ask what 'consciously deciding' actually means.

So let's get into the actual theories, because there are a few competing frameworks right now, and they're not just different ideas, they lead to completely different conclusions about things like artificial intelligence and patients in comas.

Start with Global Workspace Theory.

La Teoría del Espacio de Trabajo Global fue desarrollada por el psicólogo Bernard Baars y después expandida de forma poderosa por el neurocientífico francés Stanislas Dehaene.

Global Workspace Theory was developed by psychologist Bernard Baars and then powerfully expanded by French neuroscientist Stanislas Dehaene.

La metáfora central es la de un teatro.

The central metaphor is that of a theater.

Imagina que el cerebro procesa constantemente una cantidad enorme de información de manera inconsciente y en paralelo, como actores ensayando en la oscuridad.

Imagine the brain constantly processing a huge amount of information unconsciously and in parallel, like actors rehearsing in the dark.

La conciencia ocurre cuando esa información accede a un espacio de trabajo global y se transmite de forma amplia a toda la red cerebral, como si un actor saliera al escenario iluminado.

Consciousness happens when that information accesses a global workspace and is broadly broadcast across the brain network, as if an actor stepped into the lit spotlight.

Lo que está en el escenario, eso es lo que eres consciente.

What's on the stage, that is what you are conscious of.

So consciousness, in that model, is essentially about access and broadcasting.

Information becomes conscious when enough of the brain can talk to itself about it.

Exacto.

Exactly.

Y la teoría rival, la Teoría de la Información Integrada de Giulio Tononi, parte de un punto completamente distinto.

And the rival theory, Giulio Tononi's Integrated Information Theory, starts from a completely different point.

Tononi no pregunta dónde se difunde la información, sino cuánto la integra el sistema.

Tononi doesn't ask where information is broadcast, but how much the system integrates it.

Propone una medida matemática llamada phi, que cuantifica el grado en que un sistema genera más información como un todo que la suma de sus partes.

He proposes a mathematical measure called phi, which quantifies the degree to which a system generates more information as a whole than the sum of its parts.

Cuanto mayor sea el phi de un sistema, más consciente es.

The higher a system's phi, the more conscious it is.

Es una teoría enormemente ambiciosa porque pretende ser universal: se aplica al cerebro humano, a un gusano, a un termostato.

It's an enormously ambitious theory because it claims to be universal: it applies to the human brain, to a worm, to a thermostat.

These two theories are genuinely in conflict.

And what's remarkable is that a group of scientists actually set up what they called an adversarial collaboration to test them against each other.

In the open.

Pre-registered predictions, public results.

You don't see that often in science.

El proyecto Cogitate.

The Cogitate project.

Los resultados se publicaron en 2023 y fueron, cuando menos, incómodos para todos.

The results were published in 2023 and were, to say the least, uncomfortable for everyone.

Ninguna de las dos teorías salió indemne.

Neither theory came out unscathed.

La evidencia no respaldó de forma clara las predicciones de la Teoría de la Información Integrada sobre dónde deberían encontrarse los correlatos de la conciencia, concretamente en las cortezas posteriores del cerebro.

The evidence did not clearly support Integrated Information Theory's predictions about where the correlates of consciousness should be found, specifically in the brain's posterior cortices.

Pero tampoco confirmó plenamente las predicciones de la Teoría del Espacio de Trabajo Global.

But it didn't fully confirm Global Workspace Theory's predictions either.

Lo cual, paradójicamente, es un resultado honesto y valioso.

Which, paradoxically, is an honest and valuable result.

Science doing what science should do and finding out it doesn't know as much as it thought.

There's a third framework I want to bring in, which is predictive processing.

Karl Friston's work.

Because it reframes the whole question in a way I find genuinely disorienting.

Friston y su idea de la energía libre es uno de los marcos teóricos más influyentes y también más difíciles de la neurociencia contemporánea.

Friston and his free energy principle is one of the most influential and also most difficult theoretical frameworks in contemporary neuroscience.

La intuición central es que el cerebro no es un receptor pasivo de información del mundo exterior, sino una máquina de predicciones activa.

The central intuition is that the brain is not a passive receiver of information from the outside world, but an active prediction machine.

Constantemente genera modelos del mundo y de sí mismo, y lo que llamamos percepción no es lo que realmente está ahí fuera, sino la mejor hipótesis del cerebro sobre lo que está causando las señales sensoriales que recibe.

It constantly generates models of the world and of itself, and what we call perception is not what is actually out there, but the brain's best hypothesis about what is causing the sensory signals it receives.

La conciencia, en este marco, sería algo parecido al modelo que el cerebro tiene de sí mismo.

Consciousness, in this framework, would be something like the brain's model of itself.

Which means the self, the thing that feels like it's having the experience, is itself a prediction.

A story the brain tells about its own processes.

That's either profound or terrifying, I haven't decided.

Probablemente las dos cosas.

Probably both.

Pero hay evidencia experimental que apunta en esta dirección.

But there is experimental evidence pointing in this direction.

Los famosos experimentos con el cerebro dividido de Roger Sperry y Michael Gazzaniga en los años sesenta y setenta mostraron algo perturbador: pacientes a los que se les había seccionado el cuerpo calloso, el puente que une los dos hemisferios, funcionaban en ciertos contextos como si tuvieran dos centros de conciencia separados.

Roger Sperry and Michael Gazzaniga's famous split-brain experiments in the sixties and seventies revealed something disturbing: patients whose corpus callosum, the bridge connecting the two hemispheres, had been severed, functioned in certain contexts as if they had two separate centers of consciousness.

Y cuando el hemisferio izquierdo no sabía por qué el cuerpo había actuado de una forma concreta, inventaba una explicación.

And when the left hemisphere didn't know why the body had acted in a certain way, it invented an explanation.

De inmediato.

Immediately.

Sin vacilar.

Without hesitating.

The left hemisphere as a compulsive narrator.

Gazzaniga called it the interpreter module.

The brain doesn't report experience, it constructs a coherent story about experience after the fact.

Which makes you wonder what the story is leaving out.

Y esa pregunta tiene consecuencias prácticas enormes, no solo filosóficas.

And that question has enormous practical consequences, not just philosophical ones.

Porque si la conciencia puede separarse, si puede disociarse parcialmente de la actividad cerebral, entonces la pregunta de quién está consciente y quién no se vuelve muchísimo más difícil de responder.

Because if consciousness can be split, if it can partially dissociate from brain activity, then the question of who is conscious and who is not becomes vastly harder to answer.

Y eso nos lleva directamente al trabajo de Adrian Owen.

And that takes us directly to Adrian Owen's work.

Adrian Owen is a neuroscientist at Western University in Canada, and what he did in 2006 changed how medicine thinks about patients in vegetative states.

He asked a woman who had been diagnosed as vegetative to imagine playing tennis while she was inside an fMRI scanner.

And her brain lit up in exactly the pattern you'd expect from a healthy, conscious person imagining the same thing.

Es uno de los momentos más impactantes de la neurociencia moderna.

It is one of the most impactful moments in modern neuroscience.

Owen fue más allá: desarrolló un protocolo para comunicarse con algunos de estos pacientes.

Owen went further: he developed a protocol to communicate with some of these patients.

Les pedía que imaginaran caminar por su casa para responder 'sí' y que imaginaran jugar al tenis para responder 'no'.

He asked them to imagine walking through their home to answer 'yes' and to imagine playing tennis to answer 'no'.

Preguntó a un paciente si sentía dolor.

He asked one patient whether he was in pain.

El paciente respondió que no.

The patient answered no.

Estaba ahí.

He was there.

Escuchando.

Listening.

Sin poder mover un músculo, pero consciente.

Unable to move a single muscle, but conscious.

The implications of that are staggering.

Because how many patients have been classified as vegetative over the decades and were actually in there, aware, unable to communicate?

The legal and ethical dimensions alone.

Owen estima que alrededor del veinte por ciento de los pacientes diagnosticados como vegetativos muestran señales de conciencia oculta cuando se les somete a estas técnicas de neuroimagen.

Owen estimates that around twenty percent of patients diagnosed as vegetative show signs of covert consciousness when subjected to these neuroimaging techniques.

Veinte por ciento.

Twenty percent.

En España, como en muchos países, el diagnóstico del estado vegetativo sigue siendo clínico en la mayoría de los hospitales, y la resonancia magnética funcional con estos protocolos sigue siendo cara y poco accesible.

In Spain, as in many countries, the diagnosis of vegetative state remains clinical in most hospitals, and functional MRI with these protocols is still expensive and poorly accessible.

Hay una brecha enorme entre lo que la ciencia puede detectar y lo que los sistemas sanitarios pueden ofrecer.

There is an enormous gap between what science can detect and what healthcare systems can offer.

And this is where the choice of theory really matters.

Because Global Workspace Theory and Integrated Information Theory give you very different predictions about which patients might be conscious.

IIT, for instance, would suggest that a highly integrated but non-broadcasting system could still have significant phi, still be conscious, even without the broadcasting signature that GWT requires.

Y hay una consecuencia filosófica todavía más radical de la Teoría de la Información Integrada que Tononi acepta sin pestañear: si el phi es la medida de la conciencia, entonces cualquier sistema con phi mayor que cero tiene algún grado de experiencia subjetiva.

And there is an even more radical philosophical consequence of Integrated Information Theory that Tononi accepts without blinking: if phi is the measure of consciousness, then any system with phi greater than zero has some degree of subjective experience.

Eso incluye animales, obviamente, pero también, en principio, ciertos circuitos electrónicos muy integrados.

That includes animals, obviously, but also, in principle, certain highly integrated electronic circuits.

Lo cual es una versión contemporánea del panpsiquismo, la idea de que la mente, o algo así como la mente, es una propiedad fundamental del universo.

Which is a contemporary version of panpsychism, the idea that mind, or something like mind, is a fundamental property of the universe.

Panpsychism.

Five years ago you would have been laughed out of a serious philosophy department for saying that.

Now Philip Goff at Durham is writing bestselling books on it, David Chalmers takes it seriously.

Something shifted.

Lo que cambió es que el problema difícil resultó ser increíblemente resistente a todas las demás soluciones.

What changed is that the hard problem turned out to be incredibly resistant to all other solutions.

Si aceptas que la experiencia subjetiva no puede reducirse a la física tal como la entendemos, tienes dos opciones.

If you accept that subjective experience cannot be reduced to physics as we understand it, you have two options.

O añades algo nuevo a la física, alguna propiedad fundamental relacionada con la experiencia, que es lo que hace el panpsiquismo.

Either you add something new to physics, some fundamental property related to experience, which is what panpsychism does.

O concluyes que la experiencia subjetiva es una ilusión, que no existe realmente como tal, que es solo el cerebro contándose una historia.

Or you conclude that subjective experience is an illusion, that it doesn't really exist as such, that it's just the brain telling itself a story.

Y esa segunda opción, el ilusionismo, tiene su propio conjunto de filósofos respetables, como Keith Frankish.

And that second option, illusionism, has its own set of respectable philosophers, like Keith Frankish.

Illusionism strikes me as the most unsatisfying possible answer.

Telling me that my experience of pain is an illusion while I am in pain is a peculiar kind of comfort.

Frankish respondería que no niega que el dolor ocurra, sino que niega que tenga las propiedades fenoménicas que creemos que tiene, ese carácter cualitativo específico que los filósofos llaman 'qualia'.

Frankish would reply that he doesn't deny that pain occurs, but that he denies it has the phenomenal properties we think it has, that specific qualitative character philosophers call 'qualia'.

Es una distinción muy sutil, y honestamente, yo también la encuentro insatisfactoria.

It's a very subtle distinction, and honestly, I also find it unsatisfying.

Pero es una posición coherente.

But it is a coherent position.

The AI question is sitting right under all of this, isn't it.

Because if we can't define what consciousness is, we certainly can't say whether a large language model has it or doesn't have it.

Es el elefante en la sala, y las dos grandes teorías llegan a conclusiones opuestas.

It's the elephant in the room, and the two major theories arrive at opposite conclusions.

La Teoría del Espacio de Trabajo Global sugiere que lo crucial es la arquitectura de difusión global, y los sistemas de inteligencia artificial actuales podrían, en principio, tener algo equivalente.

Global Workspace Theory suggests that what matters is the global broadcasting architecture, and current AI systems could, in principle, have something equivalent.

En cambio, la Teoría de la Información Integrada concluye que las redes neuronales artificiales, por muy complejas que sean, tienen un phi muy bajo porque son sistemas que procesan información de manera muy modular y poco integrada.

Integrated Information Theory, by contrast, concludes that artificial neural networks, however complex they are, have very low phi because they are systems that process information in a highly modular and poorly integrated way.

Según Tononi, un ordenador convencional es, en cierto sentido, menos consciente que un lombriz.

According to Tononi, a conventional computer is, in a certain sense, less conscious than an earthworm.

An earthworm.

We've written hundreds of billions of dollars of market capitalization into AI systems that might be, by the most rigorous current theory, less aware than an earthworm.

There's a headline.

El problema, claro, es que tampoco tenemos una forma fiable de medir el phi en sistemas grandes.

The problem, of course, is that we also have no reliable way to measure phi in large systems.

Es matemáticamente intratable para redes con miles de millones de parámetros.

It is mathematically intractable for networks with billions of parameters.

Así que estamos en la situación absurda de tener una medida teórica de la conciencia que no podemos calcular en los sistemas que más nos importa evaluar.

So we are in the absurd situation of having a theoretical measure of consciousness that we cannot calculate in the very systems we most want to evaluate.

La ciencia de la conciencia está llena de este tipo de callejones.

The science of consciousness is full of these kinds of dead ends.

Which brings me back to anesthesia, and the man I watched in Jakarta.

Because anesthesia might actually be the cleanest experimental window we have.

You can turn consciousness off, you can turn it back on, you can vary the dosage.

What does it tell us?

Lo que la anestesia nos dice es parcialmente contradictorio con lo que esperaríamos de las teorías.

What anesthesia tells us is partially contradictory with what we'd expect from the theories.

Los anestésicos generales no solo suprimen la actividad cortical de forma generalizada, lo que haría pensar en el Espacio de Trabajo Global.

General anesthetics don't only suppress cortical activity broadly, which would suggest the Global Workspace.

También interrumpen específicamente la comunicación entre regiones del cerebro, lo que reduce la integración de la información, consistente con la idea de la phi.

They also specifically interrupt communication between brain regions, reducing information integration, consistent with the phi idea.

Además, hay algo fascinante: estudios con electroencefalografía muestran que bajo anestesia el cerebro emite muchas más señales de las que recibe del exterior, como si se ensimismara.

Furthermore, there is something fascinating: EEG studies show that under anesthesia the brain emits far more signals than it receives from the outside, as if it turned inward.

Como si colapsara sobre sí mismo.

As if it collapsed upon itself.

The brain talking to itself and going nowhere.

And we still don't know the mechanism by which any anesthetic actually produces unconsciousness.

For something we do millions of times a year in operating rooms around the world, that ignorance is remarkable.

Eso es completamente cierto.

That is entirely true.

Los anestesiólogos trabajan con una eficacia práctica enorme y con una comprensión mecanística muy limitada.

Anesthesiologists work with enormous practical efficacy and very limited mechanistic understanding.

Y eso, si lo piensas bien, resume el estado de todo el campo.

And that, if you think about it, summarizes the state of the entire field.

Podemos hacer cosas, podemos detectar cosas, podemos medir cosas.

We can do things, we can detect things, we can measure things.

Pero la pregunta de por qué hay algo que se siente como estar despierto, en lugar de que simplemente ocurran procesos sin ninguna perspectiva interior, sigue sin respuesta.

But the question of why there is something it feels like to be awake, rather than processes simply occurring with no interior perspective, remains unanswered.

Y quizás lo más honesto que podemos decir hoy es que hemos aprendido a formular la pregunta con mucha más precisión que hace cincuenta años.

And perhaps the most honest thing we can say today is that we have learned to formulate the question with much more precision than fifty years ago.

Eso no es poco.

That is not nothing.

Four hundred years from Descartes and his pineal gland, and the achievement is that we know exactly how hard the question is.

I find that oddly encouraging rather than depressing.

The map of our ignorance got more accurate.

Un mapa de la ignorancia.

A map of ignorance.

Me gusta esa frase.

I like that phrase.

Y tiene otra dimensión que no hemos mencionado: la conciencia no es solo un problema científico.

And it has another dimension we haven't mentioned: consciousness is not just a scientific problem.

Es el fundamento de toda la ética, de todo el derecho, de toda la política.

It is the foundation of all ethics, all law, all politics.

Si no sabes qué sistemas son conscientes, no sabes qué sistemas merecen consideración moral.

If you don't know which systems are conscious, you don't know which systems deserve moral consideration.

Eso lo hace urgente de una manera que va más allá del laboratorio.

That makes it urgent in a way that goes far beyond the laboratory.

Patients, animals, AI systems.

The question of who is in there is, in the end, the most consequential question there is.

Octavio, before we go, something you said earlier caught my ear.

You used 'conciencia' throughout, but I know in Spanish there's sometimes 'consciencia' too.

Are those the same word, or am I hearing something that isn't there?

Estás escuchando algo real, y es más complicado de lo que parece.

You're hearing something real, and it's more complicated than it looks.

En español, 'conciencia' funciona para dos conceptos que en inglés están completamente separados.

In Spanish, 'conciencia' does the work of two concepts that in English are completely separate.

Por un lado, es la conciencia moral, lo que en inglés es 'conscience', ese sentido interno de lo que está bien y lo que está mal.

On one hand, it is moral conscience, what in English is 'conscience', that internal sense of right and wrong.

Por otro lado, es la conciencia en el sentido que hemos estado discutiendo hoy, la experiencia subjetiva, lo que en inglés es 'consciousness'.

On the other hand, it is consciousness in the sense we've been discussing today, subjective experience, what in English is 'consciousness'.

La misma palabra carga con los dos significados, y el contexto es lo que distingue cuál aplica.

The same word carries both meanings, and context is what tells you which one applies.

So when a Spanish speaker says someone 'tiene mucha conciencia', they might mean the person has a strong moral compass, or that the person is highly self-aware.

English had to invent a whole separate word to split those apart.

'Consciencia' con esa 's' extra existe también, y en contextos formales o académicos se suele reservar para el sentido psicológico y filosófico, lo que haría el equivalente directo de 'consciousness'.

'Consciencia' with that extra 's' also exists, and in formal or academic contexts it tends to be reserved for the psychological and philosophical sense, making it the direct equivalent of 'consciousness'.

Pero en el habla cotidiana la distinción se pierde, y la mayoría de los hispanohablantes usa 'conciencia' para todo.

But in everyday speech the distinction dissolves, and most Spanish speakers use 'conciencia' for both.

Lo interesante es que esa ambigüedad quizás no sea un defecto del idioma.

What's interesting is that this ambiguity is perhaps not a flaw in the language.

Quizás refleja una intuición antigua de que el conocimiento de uno mismo y la moral no son dos cosas separadas, sino la misma cosa vista desde ángulos distintos.

Perhaps it reflects an ancient intuition that self-knowledge and morality are not two separate things, but the same thing seen from different angles.

Self-awareness and conscience as one word.

Descartes spent his career trying to separate the mind from the body, and Spanish never bothered to separate the mind from its own moral sense.

I'm going to think about that for a while.

Probably next to a glass of wine, no ice.

Sin hielo, exactamente.

No ice, exactly.

Hay esperanza para ti todavía, Fletcher.

There's hope for you yet, Fletcher.