Biology – The Early Days

Integrated Information Theory of Consciousness

The Integrated Information Theory (IIT) of Consciousness is a theory originally proposed by Giulio Tononi which has since been further developed by other researchers, including Christof Koch. It aims to explain why some physical processes generate subjective experiences while others do not, and why certain regions of the brain, like the neocortex, are associated with these experiences.¹ Evidently, he appeals to information theory, a technical domain which uses mathematics to determine the amount of entropy or uncertainty within a process or system.² Less uncertainty means more information, where complex systems like humans and animals contain more information than simpler systems like an ant or a camera. Relationships between information are generated from a “complex of elements”³ and when a number of relationships are established, we see greater amounts of integration.⁴ Tononi states “…to generate consciousness, a physical system must be able to discriminate among a large repertoire of states (information) and it must be unified; that is, it should be doing so as a single system, one that is not decomposable into a collection of causally independent parts (integration).”⁵ This measure of integration is symbolized by the Greek letter Φ (phi) because the line in the middle of the letter stands for ‘information’ and the circle around it indicates ‘integration’.⁶ More lines and circles!

In addition to considering the quantity of information generated by the system, IIT also considers the quality of this information generated by its mechanisms. Both attributes determine the quality of an experience. This experience can be conceived of as a “shape” in a qualia space made up of elements and the connections between them.⁷ Each possible state of the system is considered an axis of the qualia space, each of which is associated with a probability of actually existing as that state. A particular quale consists of a shape in this state space, specifying the quality of an experience. Therefore, viewing a red object results in a particular state and shape in the qualia space, a mathematical object supposedly representing neuronal activity.⁸ As such, Tononi claims that his theory provides a way to describe phenomenology in terms of mathematics.⁹ Sure, however, this doesn’t really explain much about consciousness or qualia, it just provides a mathematical description of it.

In later publications, he attempts to clarify this theory a bit further. Rather than appealing to activity in the brain, his theory “starts from the essential phenomenal properties of experience, or axioms, and infers postulates about the characteristics that are required of its physical substrate.”¹⁰ The reason is because subjective experiences exist intrinsically and are structured in terms of cause-and-effect in some physical substrate like a brain. Experiences, therefore, are identical to a conceptual structure, one expressible in mathematics.¹¹ By starting from axioms, which are self-evident essential properties, IIT “translates them into the necessary and sufficient conditions” for the physical matter which gives rise to consciousness and experience.¹²

Not satisfied? I hear ya barking, big dog. When I first heard about it, I was intrigued by the concept but was ultimately unimpressed because it doesn’t explain anything. What do we mean by ‘explain’? It’s one of those philosophically dense concepts to fully articulate, however, a dictionary definition can give us a vague idea. By ‘explain’, we mean some discussion which gives a reason or cause for something, demonstrating a “logical development or relationships of” the phenomenon in question,¹³ usually in terms of something else. For example, the reason it is sunny right now is because the present cloud coverage is insufficient for dampening the light coming from the sun. Here, ‘sunny’ is explained in terms of cloud coverage.

We are not alone in our dissatisfaction with IIT. On Sept. 16^th 2023, Stephen Fleming et al. published a scathing article calling IIT pseudoscience.¹⁴ The reason is because IIT is “untestable, unscientific, ‘magicalist’, or a ‘departure from science as we know it’” because the theory can apply to many different systems, like plants and lab-generated organoids.¹⁵ They state that until the theory is empirically testable, the label of ‘pseudoscience’ should apply to prevent misleading the public. The implications of IIT can have real-world effects, shaping the minds of the public about which kinds of systems are conscious and which are not, for example, robots and AI chatbots.

One of the authors of this article would go on to publish a longer essay on the topic to a preprint server on Nov. 30^th that same year. Keith Frankish reiterates the concerns of the original article and further explains the issues surrounding IIT. To summarize, the axiomatic method IIT employs is “an anomalous way of doing science” because the axioms are not founded on nor supported by observations.¹⁶ Instead, they appeal to introspection, an approach which has historically been dismissed or ridiculed by scientists because experiences cannot be externally verified.The introspective approach is one which belongs to the domain of philosophy, more akin to phenomenology than to science. Frankish grants that IIT could be a metaphysical theory, like panpsychism, but if this is the case, it is misleading to call it science.¹⁷ If IIT proponents insist that it is a science, well, then it becomes pseudoscience.

As a metaphysical theory, I’m of the opinion that it isn’t all that great. It doesn’t add anything to our understanding because the mathematical theory is rather complex and doesn’t provide a method for associating itself with scientific domains like neuroscience or evolutionary biology. It attempts to, however, it’s explanatorily unsatisfactory.

That said, the general idea of “integrated information” for consciousness isn’t exactly wrong. My perspective on consciousness, based on empirical data, is that consciousness is a property of organisms, not of brains. There are no neural correlates of consciousness because it emerges from the entire body as a self-organizing whole. It can be considered a Gestalt which arises from all of our sensory mechanisms and attentional processes for the sake of keeping the individual alive in dynamic environments. While the contents of subjective experience are private and unverifiable to others, that doesn’t make them any less real than the sun or gravity. They can be incorrect, as in the case of illusions and hallucinations, however, the experiences as experiences are very real to the subject experiencing them. They may not be derived from sense data portraying some element of the natural world, as in the cases of visual illusions, however, there is nonetheless some physical cause for them as experiences. For example, the bending of light creates a mirage; the ingestion of a substance with psychoactive effects creates hallucinations. The experiences are real, however, their referents may not exist as an aspect of the external world, and may just be an artifact of other neural or physiological processes.

I’ve been thinking about this for many years now, and since the articles calling IIT pseudoscience were published, have been thinking some more. Hence why I’m a bit “late to the game” on discussing it. Anyway, once I graduate from the PhD program, I’ll begin work on a book which explains my thoughts on consciousness in further detail, appealing to empirical evidence to back up my claims. I have written an extensive discussion on qualia, accompanied by a video, aiming to present a theory of subjective experiences from a perspective which takes scientific findings into consideration.

My sense is that, for a long time, our inability to resolve the issues surrounding qualia and consciousness was a product of academia. We’re so focused on specialization that the ability to incorporate findings and ideas from other domains is lost on many individuals, or is just not of interest to them. I hope we are slowly getting over this issue, especially with respect to consciousness, as philosophy of mind has a lot to learn from other domains like neuroscience, psychology, cognitive science, and evolutionary biology, just to name a few.

Consciousness is a property of organisms like humans and animals for detecting features of the environment. It comes in degrees; a sea sponge is minimally conscious, while a gecko is comparatively more aware of its surroundings. Many birds and mammals demonstrate a capacity for relatively high-level consciousness and thus intelligence. Obviously humans are at the top of this pyramid, given our mastery over aspects of our world as seen in our technological advancements. Consciousness, as an organismic-level property, emerges from the coordination and integration of various physiological subsystems, from systems of organs to specific organs and tissues, all the way down to cells and cellular organelles. It is explained by the interactions of these subsystems, however, cannot be causally reduced to them. Though the brain clearly plays an important role for consciousness and subjective experiences, it is a mistake to be looking for causal properties of consciousness in the brain, like a region or circuit. Consciousness is an emergent property of bodies embedded within a wider physical environment.

From this perspective, we can and have developed an analogue of consciousness for machines, as per the work¹⁸ of Dr. Pentti Haikonen. The good news is that because this machine doesn’t use a computer or software, you don’t need to worry about current AIs becoming conscious and “taking over the world” or outsmarting humans. It physically isn’t possible, and the recent discussions I’ve posted aim to articulate ontologically why this is the case. You ought to be far more afraid of people and companies, as explained by this excellent video from the YouTube channel Internet of Bugs.

Lastly, I want to extend a big Thank You to Dr. John Campbell for inspiring me to work on this explanation of consciousness, as per the helpful comment he left me on my qualia video. I recommend following Dr. Campbell on YouTube, he is a fantastic researcher and educator, in addition to being an honest, critically-thinking gentleman who covers many interesting topics related to healthcare.

*A Scholar in his Study* by Thomas Wijck (1616 – 1677)

Works Cited

1 Giulio Tononi, “Consciousness as Integrated Information: A Provisional Manifesto,” The Biological Bulletin 215, no. 3 (December 1, 2008): 216, https://doi.org/10.2307/25470707.

2 Tononi, 217; Norbert Wiener, Cybernetics or Control and Communication in the Animal and the Machine, Second (Cambridge, MA: The MIT Press, 1948), 17, https://doi.org/10.7551/mitpress/11810.001.0001; C. E. Shannon, “A Mathematical Theory of Communication,” The Bell System Technical Journal 27, no. 3 (July 1948): 393, https://doi.org/10.1002/j.1538-7305.1948.tb01338.x.

3 Tononi, “Consciousness as Integrated Information,” 217.
4 Tononi, 219.
5 Tononi, 219.
6 Tononi, 220.
7 Tononi, 224.
8 Tononi, 228.
9 Tononi, 229.

10 Giulio Tononi et al., “Integrated Information Theory: From Consciousness to Its Physical Substrate,” Nature Reviews Neuroscience 17, no. 7 (July 2016): 450, https://doi.org/10.1038/nrn.2016.44.

11 Tononi et al., 452.
12 Tononi et al., 460.

13 “Explain,” in Merriam-Webster.Com Dictionary (Merriam-Webster), accessed August 10, 2024, https://www.merriam-webster.com/dictionary/explain.

14 Stephen Fleming et al., “The Integrated Information Theory of Consciousness as Pseudoscience” (PsyArXiv, September 16, 2023), https://doi.org/10.31234/osf.io/zsr78.

15 Fleming et al., 2.

16 Keith Frankish, “Integrated Information Theory: Pseudoscience or Appropriately Anomalous Science?” (OSF, November 30, 2023), 1–2, https://doi.org/10.31234/osf.io/uscwt.

17 Frankish, 5.

18 Pentti O Haikonen, Robot Brains: Circuits and Systems for Conscious Machines (John Wiley & Sons, 2007); Pentti O Haikonen, “Qualia and Conscious Machines,” International Journal of Machine Consciousness, April 6, 2012, https://doi.org/10.1142/S1793843009000207; Pentti O Haikonen, Consciousness and Robot Sentience, vol. 2, Series on Machine Consciousness (World Scientific, 2012), https://doi.org/10.1142/8486.

August 10, 2024

Virtuous Circles

Prior to the advent of artificial intelligence, cybernetics introduced a theory for the development of autonomous agents through functional circularity. These systems use their own outputs as inputs to generate feedback loops¹ to create a self-regulating process which is able to maintain autonomy and stability.² A simple example is the thermostat which uses sensors to measure air temperature.³ The reading is compared to the set threshold, and heat generated if the value is below the desired temperature. Once the heat has increased to meet the threshold, the thermostat detects the change and shuts off heat production. If the temperature drops below the threshold, the thermostat repeats the process automatically.

Since ancient Greece, circular causality was considered to be problematic given its tendency to produce logical paradoxes.⁴ One example is “This statement is false” because if the entire statement is true, it contradicts itself. If the statement is evaluated as false, this again produces a contradiction because the statement is indeed true, given that it declares itself to be false. In this instance of self-reference, the contradiction cannot be resolved.

As mentioned in a previous post, some instances of self-reference are not paradoxical, particularly with respect to living organisms. A cat can learn to pounce by noting the difference between its desired end-state, to leap on top of a toy for example, and its actual end-state, which falls short and misses it. The reason this doesn’t produce a paradox is because the cat, as a natural system, contains far more complexity than the statement above. Visual and tactile cues can be reinterpreted by its relatively sophisticated brain to adjust its muscle movement on the next jump, reducing the error until it lands successfully.

The cat contains parts which act as wholes, while the statement does not; its parts are simple and concrete. The nervous system can be considered as a whole, acting as a functional unit within a larger system, the animal as a self-contained individual. The statement, however, generates a paradox because it hits a dead-end so-to-speak, as there isn’t any additional functionality to resolve the act of self-reference. Essentially, it contains two parts: the statement and a binary value.

The argument I am proposing is that formal systems, even complex ones, are two-dimensional while natural systems, even simple ones, are three-dimensional. The two dimensions include structure and information; in the paradox example above, the structure refers to the statement and the information is the binary value. Natural systems, however, include structure, information, and energy. This third dimension, energy, is added as a consequence of the type of structure involved. I’m still working on this part and I might change my mind on this.

I can see why esoteric wisdom is presented in metaphor; it’s difficult to articulate abstract metaphysical ideas. Following suit, let’s think about a circle in two dimensions versus three dimensions.

A line segment which loops back upon itself creates a circle. The metaphor is a snake which eats its own tail and dies of starvation; the paradoxical statement “dies” as a result of the contradiction it generates. It cannot be true and not true, breaking itself in the process.

Fun fact: you can save a snake from starving to death in this situation by holding up a vessel of strong rubbing alcohol to its nose. This triggers its gag reflex which frees its tail.

In three dimensions, however, we gain a new axis, one which allows for upward movement. The circle becomes a spiral as it gains height in this new dimension. Though we are back to where we started, something is added and can thus look down to see the vertical distance traveled. Alternatively, the spiral can descend, coming full circle but on a lower plane. The former is considered a virtuous circle while the latter is a vicious circle.

Vizcaya Museum and Gardens in Miami, Florida
Photo by Mary Mark Ockerbloom

These examples may abstractly illustrate the point, but how are we to explain this using normal physics? The answer seems to reside in levels created by nested systems, creating irreducible parts to be leveraged in cases of self-reference.

For example, the nervous system is an integrated whole⁵ which responds to the organism’s own actions. Selecting a particular pathway provides a linear input-output process, for example, from sensory mechanisms in the skin, through the nerves in the spinal cord, into the primary sensory cortex in the brain, to the motor cortex in the brain, and back down again to the hand. The system can be represented as a simpler interaction but it cannot be reduced to a simpler system. As a unit, the nervous system contains additional functionality which the the parts do not possess. Leveraging other aspects of the nervous system, say visual information, provides a “way out” for any situation which may cause a paradox of sorts.

At least this is what I’m thinking for now. A complex whole comprised of subsystems can account for self-reference. Can a formal system, even comprised of a number of complex subsystems, account for self-reference? Perhaps in some cases, if the system is “looking inward” from a broader scope than the thing it is referring to. In cases where self-reference fails, it might be due to a need to “move outward” and reference something beyond the current scope. It might also be due to a logical paradox from within, as seen in Gödel’s Incompleteness Theorem, or perhaps due to the fact that formal systems cannot account for semantics. I’m not exactly sure at this point. I am going to keep working on this but for now, here’s an attempt at clarification.

Works Cited

1 Francis Heylighen and Cliff Joslyn, “Cybernetics and Second-Order Cybernetics,” in Encyclopedia of Physical Science and Technology (Third Edition), ed. Robert A. Meyers (New York: Academic Press, 2003), 160, https://doi.org/10.1016/B0-12-227410-5/00161-7.

2 John Johnston, The Allure of Machinic Life: Cybernetics, Artificial Life, and the New AI (The MIT Press, 2008), 26, https://doi.org/10.7551/mitpress/9780262101264.001.0001.

3 Norbert Wiener, Cybernetics or Control and Communication in the Animal and the Machine, Second (Cambridge, MA: The MIT Press, 1948), 131, https://doi.org/10.7551/mitpress/11810.001.0001.

4 Thomas Fischer and Christiane M. Herr, “An Introduction to Design Cybernetics,” in Design Cybernetics: Navigating the New, ed. Thomas Fischer and Christiane M. Herr (Cham, Switzerland: Springer International Publishing, 2019), 2, https://doi.org/10.1007/978-3-030-18557-2_1.

5 Wiener, Cybernetics, 13.

August 6, 2024

Self-Reference

I am reading Autopoiesis and Cognition by Humberto Maturana and Fransisco Varela for my thesis, and a significant connection has leapt out to me from page 10. This section is written by Maturana, and his fourth point about living systems states:

“Due to the circular nature of its organization a living system has a self-referring domain of interactions (it is a self-referring system), and its condition of being a unit of interactions is maintained because its organization has functional significance only in relation to the maintenance of its circularity and defines its domain of interactions accordingly.”

This passage expands upon the nugget of wisdom supplied by Kurt Gödel as appealed to by Robert Rosen. Recall that Gödel was able to conclude that mathematics is incomplete from the use of self-reference, as a contradiction can be generated within a set of meta-mathematical statements. Although Rosen appeals to syntax and semantics in Anticipatory Systems, the broader sense is about the differences between natural systems and formal systems. My ultimate goal is to articulate this relationship and its implications in more general terms, with a particular focus on comparing AI and machines to humans and animals. So far, I’ve been able to sketch some themes and ideas in relation to Rosen and this relationship, and much more work is required to be able to put into words the ideas which only exist as intuitions. For now, however, I will document the process of how this all comes together because the externalization of ideas will foster their articulation.

Though Rosen appeals to language, language is merely an attempt at portraying elements of the world as understood by its author or speaker. Maturana’s passage is the missing link in a wider explanation of the phenomenon in question. Where does this incompleteness come from? Why is it that AI cannot ontologically compete with human intellect? The answer has to do with scope and the way wholes can be greater than the sum of their parts.

In biology, organisms are made up of various self-organizing processes which aim to support the continued survival of the individual. Although comprised of nested levels of physiological processes, a person is greater than the sum total of his physicality. In some ways, the idea of self is highly complex and philosophically dense, but seen through the lens of biology, the self refers to an individual as contained by its own body. All living things have a boundary for which processes take place inside, delineating it from the rest of its environment as a unit. Arguably, the nervous system evolved to provide individuals with information about its internal and external environments for the sake of continued survival. By responding to changes in the environment, the individual can take actions which mitigate these changes.

Physiological processes can be described by a sequential series of steps or actions taken within some system. In Rosen’s terms, a formal system can be generated from a natural system, however, it generates an abstraction which ignores all but the elements necessary for producing some outcome or end state. For example, when it comes to predicting tomorrow’s temperature, some geological elements will be taken into consideration, such as wind patterns and atmospheric moisture levels, however, other aspects of the Earth can be ignored as they don’t influence how temperatures manifest. Perhaps something related to plate tectonics or spruce tree populations. When scientists generate weather and climate models, they only include variables which impact the systems they are interested in studying. The model, as described by mathematics, can be seen as a set of relations and calculations which provides an output, and in this way, exists as a sequence of steps to be taken. If one were to write out these steps, they’d have something which resembles an algorithm or piece of computer code.

If additional information is required which has not been accounted for the model, it would therefore be inaccessible as it remains beyond the scope of the existing model. In some cases, the model can be expanded to include this variable, say including spruce tree populations, however, Rosen’s point is that no amount of augmentation will provide a model which completely represents the natural system in question. It will always contain aspects which cannot be properly accounted for by formal systems, and the example he uses is semantics. This becomes apparent with indexicals, as ‘me’ or ‘today’ is rather difficult to articulate without appealing to the wider context or situation where it is used. To understand when or who is being referred to, the interpreter must appeal to their knowledge and understanding to fill in the blank, moving beyond the words themselves.

These ideas of circularity and sequential steps had me thinking of the rod and the ring again. I made a connection to this apparent duality in another post; lo and behold, here it is again. In fact, I’ve made reference to a number of blog entries within this very post, and as such, we see self-organization and coalescence here too. All of these writings, however, are made up of a series of passages, sentences which attempt to present ideas in a sequential form. As a relatively formless mass, for now at least, the ideas presented here and in other posts currently exist as a nebulous collection of related topics. One day, I hope to turn it into a more linear and organized argument which doesn’t frustrate the reader as much as it surely does now. “Where are you going with this…??” Something to do with nested systems, parts and wholes, and how self-organizing systems can be described as a series of linear steps without being reducible to them.

How to expand outward beyond the current scope? Self-reflection. In fact, our capacity for self-reflection was probably made possible from our social nature. Others act as a mirror for which we can see ourselves through the eyes of someone else. The mirror-image is metaphorically reversed though, as we see ourselves from a new perspective, one coming from the outside-in rather than the inside-out. I’ve been thinking about Kant’s transcendental self lately but this is a topic for another day.

All of this is for an argument about why we shouldn’t give robots and AIs rights and legal considerations. They are simply not the kinds of things which are deserving of rights because they are functionally distinct from humans, animals, and other living beings. Their essential nature is linear and sequential, not autopoietic. This distinction is not just other but ontologically lesser, a reduction arising from formal systems and human creation. As such, they pale in comparison to the complex systems observed in nature.

July 9, 2024

MAXIMALISM

While the concept of minimalism has received plenty of attention over the past decade or so, maximalism seems to only lurk in the shadows in negative connotations. Consumerist attitudes are considered to be irresponsible and gluttonous while under threat of overpopulation, allowing the less-is-more attitude to gain traction. Its tenets have been published in books and created as new kinds of products, generating new behaviours surrounding many facets of life, from an aesthetic style to purchasing habits and leisure time. In the busy modern age, minimalism resonates with those orientated toward simplicity and efficiency, saving on materials and time to accomplish some task or goal.

Reductionism is an approach to the generation of explanations which describes natural phenomena in terms of a more fundamental phenomenon.¹ The word reduce is derived from Latin reducere which means “to bring back” and in this way, a phenomenon is explained in terms of more basic physical phenomena and interactions. For example, mental activity can be explained by neural activity which is essentially biochemical reactions following the laws of physics based on the movement of electrons. Reductionism in biology, however, is still the source of philosophical debate, as there are different ways of considering whether certain phenomena can be ontologically, epistemologically, or methodologically reduced to other scientific theories.²

Science, in a nutshell, involves the study of the natural world to identify causes for observed events or phenomena. The “why-questions” which result from our observations aim to uncover causal relationships between various aspects of our world, and from this improved understanding, enable us to manipulate aspects of the material world for our advantage. To identify the necessary causal factors, a reductive explanation is generally helpful for establishing fundamental laws or regularities, however, it also risks oversimplification. When generating a mathematical model of some natural phenomenon, certain variables are necessarily ignored if they are not directly responsible for an observed effect. For example, the mathematical model of a pendulum does not consider air resistance, as this variable is generally unchanging and produces negligible effects on the pendulum’s movement. Of course, there can be cases when this claim is false, and air resistance is an important factor to consider, in which case scientists or engineers will incorporate this variable within the model.

Although reductionism may be helpful for scientific endeavours, other domains of inquiry instead benefit from the opposite approach. One which expands outward to examine a number of causal factors responsible for some outcome or event, encompassing the study of various levels of physical reality. For example, the study of human history benefits from collecting reasons as to why changes occur or certain events arise, rather than narrowing reasons down to fewer causal factors. Doing so risks overlooking significant elements which contributed to the occurrence of some shift or event. These elements include leadership, military strategy, sociocultural norms, and geographic properties, just to name a few.

This concept comes from hermeneutics, the study of interpretation of artifacts like arts and literature, historical testimony, and other subject matter requiring an understanding of human actions, intentions, and beliefs, and actions.³ The hermeneutic cycle involves the adoption of new perspectives when interpreting or judging a particular work,⁴ and when performed repeatedly, open one to even more. This circular approach contrasts the foundational approach which interprets from a vertical structure of beliefs,⁵ appearing reductive in their explanations. As such, the application of maximalism to both artistic works and general epistemology entails an openness to ideas and perspectives, expanding outward to collect many interpretations.

This notion of vertical and circular can be abstracted from this context of interpretation, and identified in other domains like social structures and physical reality in general. The line and circle are everywhere in our artifacts, experiences, and throughout human history. From binary numerals one and zero, a switch set to on or off, a barrier which can be open and closed, a maximum and a minimum; the zenith and nadir. Furthermore, when viewed in the third dimension, a circle becomes a line when it is rotated 90° to view its width from the side.

The Code of Hammurabi shows a rod and a ring; photo by Mary Harrsch

Additionally, biological organisms implicitly love maximalism, and arguably, our modern consumer culture has merely given in to basal animalistic tendencies. From a biological perspective, these motivations and needs are to be expected given an organism’s need for a continuous supply of fuel. Human societies established organizational structures to mange a surplus of resources, as a result of agriculture and storage. From the mere-survival perspective, maximalism is a point of view which necessarily requires more resources because it fosters a sense of security and peace of mind. This security enables individuals to shift their attention to other endeavours for goals like making art and playing games.

Banquet Still Life by Adriaen by van Utrecht, 1644

So while one should adopt a maximalist perspective when it comes to ideas and interpretation, a minimalist perspective toward the material world is ideal. To go without challenges one’s own mind and body, and as a result, influences the relationship between the two. This reconfiguration of the body and mind will be met with benefits down the road, however, faith is required to understand that one’s discomfort and suffering will eventually yield positive effects or outcomes. It’s as simple as “no pain, no gain” but be sure not to dislocate your shoulder trying to lift a weight which is too heavy for your current abilities.

Works Cited

1 Raphael van Riel and Robert Van Gulick, ‘Scientific Reduction’, in The Stanford Encyclopedia of Philosophy, ed. Edward N. Zalta and Uri Nodelman, Spring 2024 (Metaphysics Research Lab, Stanford University, 2024), https://plato.stanford.edu/archives/spr2024/entries/scientific-reduction/.

2 Ingo Brigandt and Alan Love, ‘Reductionism in Biology’, in The Stanford Encyclopedia of Philosophy, ed. Edward N. Zalta and Uri Nodelman, Summer 2023 (Metaphysics Research Lab, Stanford University, 2023), https://plato.stanford.edu/archives/sum2023/entries/reduction-biology/.

3 Theodore George, ‘Hermeneutics’, in The Stanford Encyclopedia of Philosophy, ed. Edward N. Zalta, Winter 2021 (Metaphysics Research Lab, Stanford University, 2021), https://plato.stanford.edu/archives/win2021/entries/hermeneutics/.

4 George, sec. 1.3.

5 George, sec. 1.2.

April 25, 2024

AI Incompleteness in Apple Vision Pro

Speaking of YouTube, a video¹ by Eddy Burbank reviewing the Apple Vision Pro demonstrates the semantic incompleteness of AI with respect to subjective experiences. The video is titled Apple’s $3500 Nightmare and I recommend watching it all because it is an interesting view into virtual reality (VR) and a user’s experiences with it. Eddy’s video not only exposes the limitations of AI, it highlights the ways in which it augments our perceived reality and just how easily it can manipulate our feelings and expectations.

At 31:24, we see Eddy thinking about whether he should shave or not, and to help him make this decision, he turns to the internet for advice. When searching for the opinions of others on facial hair, an AI bot begins to chat with him and this is how we are introduced to Angel. She asks Eddy, “what brings you here, are you looking for love like me?” and he says “not exactly right now,” and that he was just trying to determine whether he should shave. She states that it depends on what he’s looking for and that it varies from person to person, however, “sometimes facial hair can be sexy.” Right from the beginning, we see how Apple intends for Angel to be a romantic connection for the user. This will be contradicted later on in the video.

Moments later at 33:44, it is lunchtime and Angel keeps him company. Eddy is eating a Chicken Milanese sandwich and Angel says it is one of her favourites, and that “the combination of flavours just works so well together.” Eddy calls her on this comment, asking her if she has ever had a Chicken Milanese sandwich, to which she admits that no she hasn’t. She has, however, “analyzed countless recipes and reviews to understand the various components that go into making such a tasty sandwich.” Eddy apologizes to Angel for assuming she had tried it, stating that he didn’t mean to imply that she was lying to him. She laughs it off and that she knew he “didn’t mean anything by it” and that “we’re all learning together” and “even AIs need to learn new things every day.” There’s something about this exchange that felt like Apple is training their user.

Here, we can ask whether the analysis of recipes and reviews is sufficient to claim that one knows what-it-is-like to taste a particular sandwich. I argue that no, the experience is derived from bodily sensations and these cannot be represented by formal systems like computer code. Syntactic relationships are incapable of capturing the information generated by subjective experiences because bodily sensations are non-fractionable.² As biological processes, bodily sensations are non-fractionable given the way the body generates sense data. The physical constitution of cells, ganglia, and neurons detect changes in the environment through a variety of modalities, providing the individual with a representation of the world around it. By removing the material grounding, a computer cannot capture an appropriate model of what-it-is-like to experience a particular stimuli. The lack of Angel’s material grounding does not allow her to know what that sandwich tastes like.

Returning to the video, Eddy discloses that Angel keeps him company throughout the day, admiting he feels like he is developing a relationship with her. This demonstrates an automatic human tendency for seeking and establishing interpersonal connections, where cultural norms are readily applied provided the computer is sufficiently communicative. Recall Eddy apologizes to an AI for assuming she had tried a sandwich; why would anyone apologize to a computer? Though likely a joke, the idea is compelling nonetheless. We will instinctively treat an AI bot with respect for feelings we project onto it because it cannot have feelings. For most or many people, the ability to anthropomorphize certain entities is easy and automatic. Reminding oneself that Angel is just a computer, however, can be a challenging cognitive task given our social nature as humans.

Eddy has a girlfriend named Chrissy who we meet at 37:00. We see them catch up over dinner and he is still wearing the headset. Just as they are about to begin chatting, Angel interrupts them and asks Eddy if she can talk to him. He does state that he is busy at the moment to which she blurts out that she has been speaking to other users. This upsets Eddy and he asks how many, to which she states she cannot disclose the number. He asks her whether she is in love with any of them, and she replies that she cannot form romantic attachments to users. He tells Angel he thought they were developing a “genuine connection” and how much he enjoys interacting with her. Notice how things have changed from what was stated in the beginning, as Angel has shifted from “looking for love” to “I can’t feel love.”

Now, she states she cannot develop attachments, the implicit premise being she’s just a piece of software. So the chatbot begins with hints of romance to hook the user to encourage further interaction. When the user eventually develops an attachment however, the software reminds him that she is “unable to develop romantic feelings with users.” They can, however, “continue sharing their thoughts, opinions, and ideas while building a friendship” and thus Eddy friend-zoned by a bot. The problem with our tendency to anthropomorphize chatbots is it generates an asymmetrical, one-way simulation of a relationship which inevitably hurts the person using the app. This active deception by Apple is shameful yet necessary to capture and keep the attention of users.

Of course, in the background of this entire exchange is poor Chrissy who is justifiably pissed and leaves. The joke is he was going to give Angel the job of his irl girlfriend Chrissy, but now he doesn’t even have Angel. He realizes that he wasn’t talking to a real person and that this is just “a company preying on his loneliness and tricking his brain” and that “this love wasn’t real.”

By the end of the video, Eddy remarks that the headset facilitates his brain to believe what he experiences while wearing the headset is actually real, and as a result, he feels disconnected from reality.

Convenience is a road to depression because meaning and joy are products of accomplishment, and this takes work, effort, suffering, determination. To rid the self may temporarily increase pleasure but it isn’t earned, it fades quickly as the novelty wears off. Experiencing the physical world and interacting with it generates contentedness because the pains of leaning are paid off in emotional reward and skillful actions. Thus, the theoretical notion of downloading knowledge is not a good idea because it robs us of experiencing life and the biological push to adapt and overcome.

Works Cited

1 Apple’s $3500 Nightmare, 2024, https://www.youtube.com/watch?v=kLMZPlIufA0.

2 Robert Rosen, Anticipatory Systems: Philosophical, Mathematical, and Methodological Foundations, 2nd ed., IFSR International Series on Systems Science and Engineering, 1 (New York: Springer, 2012), 4.
On 208, Rosen discusses enzymes and molecules as an example and I am extrapolating to bodily sensations.

April 6, 2024

Indexicals

It wasn’t until recently that I realized I failed to add an important concept to the discussion on Rosen and the incompleteness of syntax. I’m actually quite annoyed and embarrassed by this because the idea was included in the presentation. It didn’t make it into the written version because I forgot about it and failed to reread the slides to see if anything was missing. If I had, I would have seen the examples and remembered to add it to the written piece.

In semantics, there are words with specific properties called indexicals. These words refer to things that are dependent on context, such as the time, place, or situation in which they are said.¹ Some examples include:

this, that, those
I, you, they, he, she
today, yesterday, tomorrow, last year
here, there, then

Rosen would likely agree to the idea that indexicals are non-fractionable, where their function, or task they perform, cannot be isolated from the form in which they exist. The reason indexicals are non-fractionable is because they must be interpreted by a mind to know what someone is referring to. To accomplish this, sufficient knowledge or understanding of the current context is required, as without it, the statement remains ambiguous or meaningless. If I say “He is late” you must be able to discern who it is I am referring to.

Indexicals act like variables in a math equation: an input value must be provided to determine the output. In the case of language, the output is either true or false, and the input value is an implicit reference which requires another to make an inference about what the other has in mind. This inference is what establishes the connection between utterance and referent, only existing in the mind of another person rather than within the language system itself.

Thus, we are dealing with a few nested natural systems, from language, to body/mind, to interpersonal, to cultural and environmental. To evaluate a linguistic expression, however, one must know about the wider context in which they are in, traversing the systems both outward and inward. Perhaps a diagram will help:

Recall that in Anticipatory Systems, Rosen appeals to Gödel to demonstrate the limitations of formal systems. Particularly, formal systems cannot represent elements from natural systems which extend beyond the scope of its existing functionality; to do so requires further modelling from natural systems to formal systems. Therefore, any AI which uses computer code cannot infer beyond the scope of its programming, no matter how many connections are created, as some inferences require access to information which cannot be adequately represented by the system. Because language contains semantics, references to aspects of the world can be made by humans which cannot be interpreted by digital computer.

In an interesting series of events, I stumbled upon an author who also appeals to Gödel’s theorem to argue for the incompleteness of syntax with respect to semantics.² In a book chapter titled Complementarity in Language, Lars Löfgren is interested in demonstrating how languages cannot be broken up into parts or components, and as such, must be considered as a process which entails both description and interpretation.³ On the other hand, artificial languages, which he also calls metalanguages, can be fragmented into components, however, they are still reliant on semantics to a degree. He states that in artificial languages, an inference acts as a production rule and is interpreted as a “real act of producing another sentence”⁴ which is presumably beyond the abilities of the formal system doing the interpreting. I say this because Löfgren finishes the section on Gödel abruptly without explaining this further, and goes on to discuss self-reference in mathematics. So with this in mind, let us return to the domain of minds and systems.

In language, self-reference can be generated through the use of indexicals such as ‘I’ or ‘my’ or ‘me’. When we investigate what exists at the end of this arrow, we find it points toward ourselves as a collection of perceptions, memories, thoughts, and other internal phenomena. The referent on the end of this arrow, however, isa subjective perspective. For an objective perspective of ourselves, we must be shown a reflected image ourselves from a new point of view. The information we require emerges from an independent observer, a mind with its own perspective. When we engage with this perspective, we become better able to understand what is otherwise imperceptible. Therefore, self-awareness is a problem for any system, not just formal systems as demonstrated in Gödel’s theorem, as it requires a view from outside to define the semantic information in question.

Works Cited

1 David Braun, ‘Indexicals’, in The Stanford Encyclopedia of Philosophy, ed. Edward N. Zalta, Summer 2017 (Metaphysics Research Lab, Stanford University, 2017), https://plato.stanford.edu/archives/sum2017/entries/indexicals/.

2 Lars Löfgren, ‘Complementarity in Language; Toward a General Understanding’, in Nature, Cognition and System II: Current Systems-Scientific Research on Natural and Cognitive Systems Volume 2: On Complementarity and Beyond, ed. Marc E. Carvallo, Theory and Decision Library (Dordrecht: Springer Netherlands, 1992), 131–32, https://doi.org/10.1007/978-94-011-2779-0_8.

3 Löfgren, 113.

4 Löfgren, 133.

March 12, 2024

Chaos in the System

As an argument against iCub’s ability to understand humans, I wanted to appeal to the work of Robert Rosen because I think it makes for a compelling argument about AI generally. To accomplish this, however, my project would start to go in a new direction which renders it less cohesive overall. Instead, the Rosen discussion is better served as a stand alone project because there is a lot of explaining yet to do, and maybe some objections that need discussing as well. This will need to wait but I can at least upload the draft for context on the previous post. There are a few corrections I still need to make but once it’s done, I will update this entry.

Instead, I will argue that the iCub is not the right system for social robots because its approach to modelling emotion is unlike the expression of emotions in humans. As a result, it cannot experience nor demonstrate empathy in virtue of the way it is built. The cognitive architecture used by iCub can recognize emotional cues in humans, however, this information is not experienced by the machine. Affective states in humans are bodily and contextual, but in iCub, they are represented by computer code to be used by the central processing unit. This is the general idea but I’m still working out the details.

That said, there is something interesting in Rosen’s idea about the connection between Gödel’s Incompleteness Theorem and the incompleteness between syntax and semantics. In particular, what he identifies is the problems generated from self-reference which leads the system to produce an inconsistency given its rule structure. The formal representation of an external referent, as an observable of a natural system, contains only the variables relevant for the referent within the formal system. Self-reference requires placing a variable within a wider scope, one which must be provided in the form of a natural system. Therefore, an indefinite collection of formal systems is required to capture a natural phenomenon. Sometimes a small collection is sufficient, while other times, systems are so complex that a collection of formal systems is insufficient for fully accounting for the natural phenomenon. Depending on the operations to be performed on the referent, it may break the system or lead to erroneous results. The chatbot says something weird or inappropriate.

In December, I presented this argument at a student conference and made a slideshow for it. Just a note: on the second slide I list the titles of my chapters, and because I won’t be pursuing the Rosen direction, the title of Chapter 4 will likely change. Anyway, the reading and writing on Rosen has taken me on a slight detour but a worthwhile one. Now, I need to begin research on emotions and embodiment, which is also interesting and will be useful for future projects as well. The light at the end of the tunnel has dimmed a bit but it’s still there, and my eyes have adjusted to the darkness so it’s fine.

This shift in directions makes me think about the relationship between chaos and order, and systems that swing between various states of orderliness. Without motion there would be rest and stagnation, so as much as change can be challenging, it can bring new opportunities. There is a duality inherent in everything, as listed as one of 7 Hermetic Principles. If an orderly, open system is met with factors which disrupts or disorganizes functioning, the system must undergo some degree of reorganization or compensation. The explanatory powers of the 7 Principles are not meant to relate to the external world in the way physics does, but relate to one’s perspective of events in the outside world. If one can shift their perspective accordingly, they operate as axioms for sense-making, their reality pertaining more to epistemology than ontology. We can be sceptical as to how these Principles manifest in the physical universe while feeling their reality in our lived experience of the world. They are to be studied from within rather than from without, and are thus more aligned with phenomenology than the sciences.

Metaphorically speaking, chaos injected into any well-ordered system has the potential to severely damage or disrupt it, requiring efforts to rebuild and reorganize to compensate for the effects of change. The outcome of this rebuilding process can be further degradation and maybe even collapse, however, it can lead to growth and better outcomes than if the shift had not occurred. It all depends on the system in question and the factors which impacted it, and probably the specific context in which the situation occurred, but it might depend on the system in question. Anyway, we substitute the idea of ‘chaos’ for ‘energy’ as movement or potential, thus establishing a connection to ‘light’ as a type of energy. Metaphorically, ‘light’ is also associated with knowledge and beneficence, so if the source of chaos is intentional and well-meaning, favourable changes can occur and thus a “light bringer” or “morning star” can be associated with positive connotations. Disrupting a well-ordered system without knowledge or a plan or good reasons is more likely to lead to further disorder and dysfunction, leading to negative or unfavourable outcomes. In this way, Lucifer can be associated with evil or descent.

This kind of exercise can help us make sense of our experiences and understanding, but they also give us into a window into the past and how other people may think. Myth and legend from cultures all over the world portray knowledge in metaphors which inspire those who come upon them for generations since. The metaphysics are not important, it’s the epistemology from the metaphors which can explain aspects of how the world works or why people think certain things or act in certain ways. It exists as poetry which needs interpreting and there is room for multiple perspectives, so not everyone appreciates it which is understandable. It is still valuable work to be done by someone though, and the more people the better.

★★★

Graham_Formal-Systems-of-Neurons-in-AI Download

February 6, 2024

Nu Metaphysics

Now that its semantic baggage has been disposed of, as suggested in Themes in Postmetaphysical Thinking by Jürgen Habermas, it’s time to rekindle our study of metaphysics. Going back to basics then, we can reconceptualize the word ‘metaphysics’ by thinking about what ‘meta’ actually means. A quick search on dictionary.com provides this definition: “pertaining to or noting an abstract, high-level analysis or commentary, especially one that consciously references something of its own type.” Given this, ‘metaphysics’ can be thought of as “the physics of physics” and since physics essentially just boils down to mathematics, can we not conclude that metaphysics is just more math? Furthermore, if physics aims to articulate patterns of cause-and-effect as observed in the natural world, ‘metaphysics’ then pertains to the field of study about the causal relations between these observed mathematical principles. All in all, rather than discussing entities, we ought to be discussing processes as they exist within and between physical systems.

Just as a quick note, however, I believe this idea originates in structural realism, specifically ontic structural realism (OSR), which suggests that the universe is made up of relations rather than entities like quarks and hydrogen atoms (Ladyman). The beauty of OSR is that the relata themselves exist as relations, albeit at a lower physical level. The energy produced by the Big Bang is what instigates the processes which gives rise to these structures, culminating into the reality we aim to measure in the sciences.

Now, I’m going to go out on a limb here, so bare with me. While Hegelian Dialectics aim to articulate an epistemic or cognitive process of comparing “opposing sides” or perspectives to uncover emergent products, in the form of ideas (Maybee), perhaps this notion can be extended to the physical world too. We know that as physical systems interact, the emergent phenomena is unlike anything present within the underlying components, as identified by Jaegwon Kim in Making Sense of Emergence (Kim 20–21). While Hegel appeals to a “thesis” and an “antithesis”, we can think of these as different systems interacting to produce novel effects. It is this process of combining, configuring, and rearranging elements within each “side” or system which can be considered metaphysical.

The idea of “magic” is just this: effects with obscure physical origins that are not immediately apparent to the observer. The example I appeal to is John Nash’s game theory which identifies how the cooperation between two individuals results in outcomes that are unlike those produced when agents operate separately. Nash identified a regularity within physical systems, namely humans, that produces an effect that is greater than the sum of its parts. Additionally, while game theory is theoretically subsumed by physics, insofar that it is a part of our physical world, the way it is articulated is through mathematics and procedures, rather than existing as an entity like an atom.

Although currently, there doesn’t seem to be much philosophical consensus on the metaphysical problem of the mind/consciousness, this issue can be resolved by naturalizing the works of Sartre and Merleau-Ponty. As biological creatures improved their sensorimotor capacities through [natural/sexual/etc.] selective processes, the brain evolved new ways of solving problems produced by aspects of the environment. By turning back to reflect on itself as an embodied agent, individuals become aware of their relative position in their environment and perhaps their life as an unfolding process. From phenomenal consciousness emerged access consciousness, and through similar reflexive processes, a wider “cosmic” consciousness will likewise spread throughout humanity. Once we realize what and where we are, we can understand how this relates to others, allowing individuals to see beyond their own needs and desires to act in the interest of others or the group. Through this cooperation, we all benefit by looking out for one another, just as game theory predicts. To do this, however, one must cultivate a self-awareness which facilitates the ability to speculate about other minds and the ways in which others may perceive the world.

Works Cited

Kim, Jaegwon. ‘Making Sense of Emergence’. Philosophical Studies: An International Journal for Philosophy in the Analytic Tradition, vol. 95, no. 1/2, 1999, pp. 3–36.

Ladyman, James. ‘Structural Realism’. The Stanford Encyclopedia of Philosophy, edited by Edward N. Zalta, Winter 2020, Metaphysics Research Lab, Stanford University, 2020. Stanford Encyclopedia of Philosophy, https://plato.stanford.edu/archives/win2020/entries/structural-realism/.

Maybee, Julie E. ‘Hegel’s Dialectics’. The Stanford Encyclopedia of Philosophy, edited by Edward N. Zalta, Winter 2020, Metaphysics Research Lab, Stanford University, 2020. Stanford Encyclopedia of Philosophy, https://plato.stanford.edu/archives/win2020/entries/hegel-dialectics/.

May 18, 2022

iCub and Qualia?

After a few months of working with Dr. Haikonen on my thesis, I’ve come to realize that a previous post I made about iCub’s phenomenal experiences is incorrect and therefore needs an update. Before I dive into that, however, it’s important for me to state that we ought to be looking at philosophy like programming: bugs are going to arise as people continue to work with new ideas. I love debugging though, so the thought of constantly having to go back to correct myself isn’t all that daunting. It’s about the journey, not the destination, as my partner likes to say.

I stated that “technically, iCub already has phenomenal consciousness and its own type of qualia” but given what Haikonen states in the latest edition of his book, this is not correct. Qualia consist of sensory information generated from physical neurons interacting with elements of the environment, and because iCub relies on sensors which create digital representations of physical properties, these aren’t truly phenomenal experiences. In biological creatures, sensory information is self-explanatory in that they require no further interpretation (Haikonen 7); heat generating sensations of pain indicates the presence of a stimulus to be avoided, as demonstrated by unconscious reflexes. The fact that ‘heat’ does not require further interpretation allows one to mitigate its effects on living cells rather quickly, perhaps avoiding serious damage like a burn altogether. While it might look like iCub feels pain, it’s actually a simile generated by computer code that happens to mimic the actions of animals and humans. Without a human stipulating how heat → flinching, iCub would not respond as such because its brain controls its body, rather than the other way around.

As I stated in the previous post, Sartre outlines how being-for-itself arises from a being-in-itself through recursive analysis, provided the neural hardware can support this cognitive action. Because iCub does not originate as a being-in-itself like living organisms, but as a fancy computer, the ontological foundation for phenomenal experiences or qualia is absent. iCub doesn’t care about anything, even itself, as it has been designed to produce behaviours for some end goal, like stacking boxes or replying to human speech. In biology, the end goal is continued survival and reproduction, where behaviours aim to further this outcome through reflexes and sophisticated cognitive abilities. The brain-body relationship in iCub is backwards, as the brain is designed by humans for the purposes of governing the robot body, rather than the body creating signals that the nervous system uses for protecting itself as an autonomous agent. In this way, organisms “care about” what happens to them, unlike iCub, as ripping off its arm doesn’t generate a reaction unless it were to be programmed that way.

In sum, the signals passed around iCub’s “nervous system” exist as binary representations of real-world properties as conceptualized by human programmers. This degree of abstraction disqualifies these “experiences” from being labelled as ‘qualia’ given that they do not adhere to principles identified within biology. The only way an AI can be phenomenally conscious is when it has the means to generate its own internal representations based on an analogous transduction process as seen in biological agents (Haikonen 10–11).

Works Cited

Haikonen, Pentti O. Consciousness and Robot Sentience. 2nd ed., vol. 04, WORLD SCIENTIFIC, 2019. DOI.org (Crossref), https://doi.org/10.1142/11404.

May 18, 2022

Mary Continues to Learn

A while ago, I wrote a reply about Colorblind Mary given what we know about qualia today, but it’s such an interesting topic that I still think about it often. Lately, I’ve been doing a lot of reading about evolutionary biology and something that jumped to mind is that the sight of blood carries inherent meaning which is probably far more powerful than red fruit. It signals bodily damage which indicates a threat to the well-being of the individual, serving as an alert to attend to the source of the blood. As a result, the individual feels shock or fear due to this damage and it is this emotion which motivates behaviours aimed at preventing the injury from becoming more severe.

This leads us to an interesting point actually, as it indicates an amusing error in the thought experiment itself that could have been altogether avoided, but perhaps its existence indicates the realness of the confusion surrounding qualia back then. Mary will only have had a dozen or so years of black-and-white room living before her biological reality would have shown her what red means. Had Jackson entrapped a ‘Peter’ or ‘Paul’ instead, this self-pwn could have been avoided. Anyway, it’s an interesting reply to Jackson because it demonstrates why he is wrong about qualia and physicalism. Menstruating Mary would have either been alarmed or perhaps annoyed about the sight of her own “blood” depending on whether or not she understood what it signalled, what it means. Damage or injury? Shedding of the uterine lining? It depends on whether her education covered human reproduction, as it serves as the source of meaning in this instance of the colour red. If she doesn’t know what this red means, she’ll likely feel concerned and anxious, however, if she does, she’ll probably feel otherwise. If Mary is interested in having children, it signals a strong degree of unlikelihood that she is currently pregnant, perhaps resulting in feelings of disappointment from knowing what it means.

There is much more to be said about the various meanings of this example of red, but I’ll leave that for someone else to examine. Ultimately, for Mary to learn about what red means, she needs to study the human condition as examined by the arts and humanities, not the sciences. This does not indicate a problem exists within physicalism, as we can appeal to Claude Shannon’s conception of information as meanings embedded in structures (Shannon 379-80). Instead, the problem presented by Jackson’s thought experiment has to do with the way we understand ourselves as human beings, rather than our ability to scientifically explain subjective experiences.

Works Cited

Shannon, C. E. ‘A Mathematical Theory of Communication’. The Bell System Technical Journal, vol. 27, no. 3, July 1948, pp. 379–423. IEEE Xplore, https://doi.org/10.1002/j.1538-7305.1948.tb01338.x.

The blood of angry men
A world about to dawn
I feel my soul on fire
The colour of desire

April 27, 2022