Why do AI say "I"?

 New article in French : ChatGPT et l'absence du sujet

Abstract : If there is no "subject" inside them, why do AI say "I"? 

Half the sky of Wikidata

I thought I was over with this blog where I'd not published for almost two years, but I've been back to linked data lately, through a grandfather's interest in genealogy. For what is genealogy, if not the ancestor of linked data science? Genealogical trees are maybe the first type of semantic graph ever invented. Entities (persons) linked to each other by predicates such as has father, has mother, has child, has sibling, married to, linked to places (of birth, of death, of marriage), points in time (dates of birth, marriage, death), occupations, works etc. One could think that genealogical data would be the first candidate to be exposed as linked open data. But far from it. Most genealogical data is locked in proprietary data bases, and exchanged in formats far from the semantic web standards. The largest of those data bases such as MyHeritage hold billions of records.

In the linked data world, Person is indeed the most represented type of things, but the figures are three orders of magnitude below those of the above quoted giant genealogical data silos. As I write, Wikidata contains over 4,500,000 people. The current exact value can be retrieved from this query thanks to the excellent Wikidata SPARQL interface. That other query retrieves the current number of women (declared of gender female), a little more than 700,000. A similar one yields the number of those declared as male, more than 3,000,000. It lets a number of people of which gender is neither male or female, or not specified in the data base, similar to the number of women.

Let's not nitpick on numbers, and face the obvious fact that Wikidata has a strong gender bias. Less than one person out of five in Wikidata is a woman. This is not of course a deliberate Wikidata policy, but a mirror of how the notability process works at large in our world, not only in Wikipedia (the main source of Wikidata) but also in other data sources such as library authorities. If one applies to the previous queries a supplementary filter such as people with an ISNI or VIAF identifier, the proportion stays about the same. Is this changing with time? Maybe men were more notable in old ages, and the results are more balanced nowadays. Barely. More than half of people identified in Wikidata are born after 1900, and filtering the above queries to select only people of less than 50 years (born since 1968), one finds about 200,000 women for 550,000 men. The ratio has raised up slightly over 25%. A little better, but no big deal. Not half the sky, yet.

Many women can certainly be added to Wikidata, without breaching too much the notability policy. Reading through many Wikipedia articles of so-called notable people (of either gender), one can notice that women linked to them are often quoted and named as mother, spouse, daughter, sister, with elements of description such as birth and death date, and more. But those women have not yet been considered notable enough to be the subject of a separate entry in Wikipedia, and therefore not entered in Wikidata, although often they would provide a missing genealocical link between existing elements.

What about the genealogical relationships figures? Since they are the most ancient and obvious ways of linking people, one would think they are very common in Wikidata. Far from it. Less than 10% of all people are linked, as either subject or object, by a parenthood predicate (child, mother, father, sibling, spouse). And focusing on gender again, one can find less than 15,000 mother-daughter links (declared both ways) versus more than 90,000 father-son links. The gender bias shown by the number of relationships is even more obvious than the number of entities.

Meaning, quantum process and inscrutability

The analogy between meaning and measurement in quantum mechanics is something that has been on my mind for quite a while, as attested by a couple of posts from the early years of this blog. I'm therefore walking here an old path, but with a couple of new things in mind, including a quite radical shift in my viewpoint on signification since 2005, and the current lively debate around inscrutability of machine learning algorithms. The following points sum up where I stand today.

Meaning is a process

The Web has been a wide-scale experience in applied semantics, and more and more, in applied semiotics. Our interaction with the Web is using signs, the primordial and main ones being those weird identifiers called URIs. For years, I have, with many others, struggled with the thorny issue of what those URIs actually identify, or denote, or mean, or represent, spent hours in endless debates with Topic Maps and Semantic Web people to figure the difference or similarity between subjects and topics of the former, and resources of the latter. Eventually fed up with those intractable ontological issues, I decided to keep definitely agnostic about them, to focus on the dynamic aspects.

To the question What does it mean? I answer now It means what it does. In other words, the meaning of a URI on the Web is a process, whatever happens when you use it. This process can be technically described and tracked. It includes query processing, client-server dialogue, content negociation and federation, distributed computing, and more and more artificial intelligence. But from the end-user viewpoint, the URI are now hidden under the hood, the interface with the Web using natural language signs like words and sentences, written or spoken, and more and more those application icons on the touchscreen of our mobile devices, simple signs bringing us back to hieroglyphs and magic symbols. Meaning on the Web is the (more and more complex) processing of (more and more simple) signs.

Is this conception of meaning specific to the Web? If one looks closely, the answer is no. Meaning of (often simple) signs outside the Web is also the result of a (often complex) process. Whatever its nature, a sign means nothing outside a process of signification. The Web has simply given us an opportunity to explore this reality in-depth because we have engineered those process, whereas outside the Web those process are given, we use them without question on a daily basis, and we are not aware of their complexity. The more complex the Web is becoming, and the simpler the signs we use to interact with it, the closer it seems to our "natural" (read : pre-Web) semiotic activity.

Meaning process is similar to quantum process

The evolution of the Web is also tackling the difficult issue of meaning in context. The process triggered by the use of a sign is almost never the same. The time of the query, the nature state of your client device, the state of the network, your user preferences, interaction history and rights of access, the content negociation ... make every other URI resolution a unique event. Among all possible meanings, only one is realized.

Here comes the analogy with quantum mechanics. Among all possible states of a system, of which probability distribution might be known with great accuracy, only one is realized in any quantum event. Before the event, the system is described as a superposition of all its possible states. The reduction of this pack of possibles to one realization is technically called collapse of the wave function.

Samely, before you sent a query using a sign, before you click your email icon application, everything is possible. You might have mail or not. Your spam filter might have trashed an important contract. Whatever happens means the collapse of all possible states, but one. This collapse process defines the meaning of the sign at the moment you use it.

The same way in natural conversation you would say "Will you pass me the bowl?" and of all the possible meanings of "bowl" in your interlocutor's mind, all will collapse to zero but the one which indicates the only bowl sitting on the kitchen's table in front of you.

Both meaning and quantum process are inscrutable, and it's OK

The inscrutability of reference has been discussed in depth by Quine in Word and Object (1960). Quine wrote mostly before our world of pervasive information networks, before the Web, and although he died at the eve of the 21st century, he did not write anything about the Web, unless I miss something. Which is too bad, because "Word and Object" in the framework of the Web, and singularly the Semantic Web, translates easily into "URI and Resource", but maybe Quine was a bit too old in the early days of the Web to apply his theories to this new and exciting field.

Therefore, unless I miss something, Quine did not address the reference in the dynamic aspect we discuss here. Reference is inscrutable because it's a process which involves each time a sign is used a very complex and (either in theory or in practice) inscrutable process. In human natural interpretation of signs, this meaning process involves several parts of our brains and perception/action systems in a way we just barely figure. The signs we send to the network are and will be processed in more and more complex and practically inscrutable ways, such as the machine learning algorithms we already see implemented in chatbots.

Quantum process have been known since about one century ago to be inscrutable, although some of its famous founders did not like this frontal attack against determinism at the very heart of the hardest of all sciences. Albert Einstein among others was a fierce opponent to this probabilistic view of the world, defended by quantum mechanics orthodox interpretation, and used a lot of time and energy to defend without success some "hidden variable theory". Inscrutability was here to stay in physics. It seems also here to stay in semiotics, and in information systems. This is a singular convergence, which certainly deserves to be further considered and explored.

[Further reading might include works by Professor Peter Bruza (Queensland University of Technology, Brisbane, Australia) such as Quantum models of cognition and decision or Quantum collapse in semantic space : interpreting natural language argumentation.]

The right tension of links

By 1990, at the dawn of the Web, Michel Serres was publishing Le Contrat Naturel (further translated into English as The Natural Contract). In this book the philosopher makes a strong and poetic evocation of those collective ventures where contracts are materialized by cords, lines, ropes, such as sailing and climbing. Those lines link people not only to each other, but to their apparatus (sails, winches, harnesses and spikes) and to the harsh natural elements with which they are engaged (wind and waves, ice and rocks). In high sea as in high mountain, in order to ensure the cohesion and security of the team, the lines need to be tightened. And, adds Serres, this tightening is not only a safeguard, it's also a condition for the line to convey information, in a way which is more immediately efficient than language in situations where you can't afford delays in appreciation of situation and decision. If the line is too slack, you do not feel the sail and the wind, you lose connection with your climbing mate. On the other hand, excessive tension means opposition and risk of breaking the line, and being tightly connected must not impede movement. Michel Serres does not mention martial arts, but in his excellent "guide for beginners" Aikido From the Inside Out, Howard Bornstein has similar thoughts in his chapter dedicated to connection. Connection has to be maintained just at the right level of tension, by feeling what he calls the point of first resistance.

When you connect like this, you become one with your partner in a very real, experiential way. When you move, your partner moves, at the same time and in the same direction. You are really one, in terms of movement. Your experience of movement is basically the same as if you were moving entirely by yourself.

Of course, understanding in theory those general principles will not make you an experienced sailor, climber or martial artist. You will have to practice and practice to get the quality of touch enabling you to keep the lines at the right tension, making everyone safe and giving you this wonderful feeling of being one with your teammates, partners, and the world around you.

Our online experience should abide by the same rules. All the links we are texting should be of the same quality as those of sailors, climbers and martial artists, enabling us to move together. In the stormy events we are facing, we need more than ever to reduce the slack in our connections. 

More things in heaven and earth

Horatio : 

O day and night, but this is wondrous strange!

Hamlet : 

And therefore as a stranger give it welcome.

There are more things in heaven and earth, Horatio,

Than are dreamt of in your philosophy.

Horatio would certainly be as bewildered as we are today by the evergrowing number and diversity of things modern science investigation keeps discovering at a steady pace. A recurrent motto in science papers and articles I stumbled upon lately is more than expected, as the following short review illustrates, traveling outwards from earth to heaven. 

New living species, both living and fossil ones, are discovered almost on a daily basis in every corner of our planet, from the soil of our backyards to the most unlikely and remote places, and more and more studies suggest there are way more to discover than we already have. But the number of living things might be dangerously challenged by the growing number of artificial ones, products of our frantic industry cluttering our homes, backyards, cities and eventually landfills.

Even if a very populated one, our small planet is just itself a tiny thing in the universe, among a growing number of siblings. The number and variety of bodies in the Solar System, as well as the distance we can expect to find them, have been growing beyond expectations. Closer to us, a survey of impacts on the Moon over seven years has yielded more events than expected based on previous models of the distribution of small bodies in the inner Solar System. Images of the solar atmosphere by the SOHO coronograph has yielded an impressive number of spectacular sungrazing comets. And missions to planets have unveiled a wealth of amazing landscapes, comforting hopes to discover life in some of them.

Beyond the exploration of our home stellar system, the discovery of thousands of exoplanets did not come as a real surprise (our star being an exception would have been a big one), but there again we begin to discover more than expected, from an earth-sized planet around the star next door to improbable configurations such as planets orbiting binary stars. Moreover, free-floating, or so-called rogue planets, not tied to any specific star, are certainly cruising throughout our galaxy, and although very few of them have so far been actually detected, due to the extreme difficulty of such observations, some studies suggest they may outnumber the "regular" planets, those orbiting a star. Regarding stars themselves, the most recent catalog contains over one billion of them, which is less than 1% of the estimated total star population of our Milky Way galaxy, while new studies tend to indicate that the number of galaxies in the observable universe is at least one order of magnitude higher than previously thought. Even exotic thingies such as merging black holes, of which detection is now possible based on the transient ripples they create on space-time (aka gravitational waves) appear to be more frequent than expected. And the universe has certainly more in store, including the infamous missing mass, dark matter of which nature remains unknown.

The sheer number of objects unfolding in the depths of space and time is well beyond the grasp of human imagination and cataloguing power, not to mention philosophy. But fortunately the modern Horatio gets a little help from his friends, the machines. The overwhelming tasks of data acquisition, gathering and consolidation, identification, classification, cataloguing, are now more and more delegated to machines. Artificial intelligence, and singularly machine learning technology is beginning to be applied to tasks such as classifying galaxies or transient events. Using such black box systems for scientific tasks is stumbling again on issues linked to inscrutability, which we addressed in the previous post. Scientific enquiry is a very singular endeavour where whatever works is not easily accepted and the use of inscrutable information systems can be arguably considered as a non-starter. 

There are more and more things indeed in heaven and earth that we know of, and we are more and more eager to accept the unknown ones we discover every day. But the ones our poor imagination might be forever unable to fathom are those new ghosts haunting our intelligent machines. Are we ready to welcome those strangers?

[Edited, following +carey g. butler's comments to strikethrough above intelligent. Let me be agnostic about the fact that machine learning systems (or whatever systems to come) are intelligent or not, because I don't know what intelligent means exactly, be it natural or artificial. The "ghostly" point here is inscrutability.]

I trust you because I don't know why

The ongoing quick and widespread development of neural networks and deep learning systems is triggering many debates and interrogations both practical and conceptual. Among various features of such systems, the most arguable ones are certainly inscrutability and fallibility. A deep learning system builds up knowledge and expertise, as natural intelligence does, by accumulation of experience of a great number of situations. It does better and better with time. But the drawback of this approach is that you can't open the box to understand how it achieves its expertise as you would do with a classical step-by-step algorithm (inscrutability), and the expertise is not 100% proof, it's bound to fail from time to time (fallibility). I've written on some philosophical aspects of those issues, and how they relate to ancient Chinese philosophy (in French here). 

A recent article in Nature entitled "Can we open the black box of AI" presents a very good review of those issues. And the bottom line of this article comforts me in the opinion that either all this debate is moot, or that it is not linked to this specific technology, and not even to any kind of technology. All the debate is to know if we can trust something we don't understand and which is, moreover, bound to fail at some point. This seems to fly in the face of centuries of science and technology development all based on understanding and control. 

Do we control and understand everything we trust? Or more exactly, do we need to understand and control before we trust? Most of the time, no. As children, we trust our parents and adult world to behave properly without understanding the why's and how's of this behavior. And if, growing up, we start trying to question those why's and how's, it might happen that for some reason we lose that trust. When I trust a friend to achieve what she promised, I won't, or a least I should not, try to control and check if she will do it or not, and how. Trust, in fact, if exactly the opposite of control. You trust because you can't afford to, or have not the technical or conceptual tools to, or simply believe it's useless, counter-productive or simply rude to understand and control.

That line of thought applies to more simple things that people. If I cross a bridge over a river, I don't check, and do not understand, most of the time, how it's built. I begin to check it if for some reason it seems poorly built, or rotten, looking like no one else has used it for ages. You trust food you eat because you trust your provider, you generally don't check the food chain again and again. You start to check when you suspect this chain to present some serious point of failure. It's not check before trusting, it's check because for some reason you don't trust anymore. The other way round is called paranoia.

Most of the time, you trust things to work safely as expected because so far they mostly did work safely. Based on experience, not logical analysis of how it works.This includes, and actually begins with, your own body and brain. Looking further at the world around you, you discover black boxes everywhere, and it's all right. Starting to check and control how they work is likely to lead you in some infinite recursion of effects and causes, and you will either reasonably stop at some point saying "well, it's gonna be all right", or pass the rest of your life lost in metaphysical and ontological mist, and fear of any action.

Let's face it. We trust before and without understanding and controlling. Every second of every day. And most of the time it's OK. Until it fails, at some point. We know that it will. We trust our body and brain in order to live, although we know they are bound to break down at some point. We are aware that things and people we trust are bound to fail once in a while. That's just how life goes. Parents have a second of distraction and a child dies crossing the street. Friends are stuck in a traffic jam, don't show up on time and miss their flight, bridges collapse in sudden earthquakes, hard drives break down, light bulbs explode, lovers betray each other ...

Despite of our awareness of such risk of failure, we keep trusting, and call this hope. Without trust we lose hope, and fall into depression and despair. This is a basic existential choice : trust and live, or try to control and understand everything, ask for total security, and despair because you can't find it. We trust each other although, and actually because, we don't know why. And knowing that each of us will eventually fail some day, if only once at this ultimate individual failure point which is called death, should make each of us more prone to forgiveness. 

Let me borrow those final words from the brand new and unexpected Nobel Prize in Literature. 

Trust yourself

Trust yourself to do the things that only you know best

...

Trust yourself

And look not for answers where no answers can be found

...

Immortality, a false good idea

Immortality is trendy. According to some so-called "transhumanists", it is the promise of artificial intelligence at short or medium term, at the very least before the end of the 21st century. Considering the current advances in this field, we are bound to see amazing achievements which will shake our very notions of identity (what I am) and humanity (what we are). If I can transfer, one piece after another, neuron after neuron, organ after organ, each and every element which makes my identity into a human or machine clone of myself, supposing this is sound in theory and doable in practice, will this duplicate of myself still be myself? The same one? Another one? And if I make several clones, which one will be the "true" one? Do such questions make any sense at all? All this looks really like just another, high-tech, version of the Ship of Theseus, and our transhumanists provide no more no less than the ancient philosophers answers to the difficult questions about permanence and identity this old story has been setting, more than two thousand years ago.

None of those dreamers seem to provide a clear idea of how this immortality is supposed to be lived in practice, if ever we achieve it. A neverending old age? Not really a happy prospect! No, to be sure, immortality is only worth it if it goes with eternal youth! And even so, being alone in this condition, and seeing everyone else growing old and die, friends, family, my children and their children, does not that amount to buying an eternity of sorrow? Not sure how long one could stand that. But wait, don't worry, our transhumanists will claim, this is no problem because just everybody will be immortal! Everybody? You mean every single one of the 10 billion people expected to be living by 2100? Or only a very small minority of wealthy happy few? But let's assume the (highly unlikely) prospect of generalized immortality by 2050. In that case it will not be 10 but 15 billion immortal people at the end of the century if natality does not abate.That's clearly not sustainable. But maybe when everyone is immortal, there will be no need to have children anymore, and maybe even at some point it will be forbidden due to shrinking resources. Instead of seeing your children die like in the first scenario, you will not see children anymore. Not sure which one is the worst prospect!
Either way, alone or all together, immortality is definitely not a good idea. And if it were, life would have certainly invented and adopted it long ago. But since billions of years, evolution and resilience of life on this planet despite all kinds of cataclysms (the latest being humanity itself) is based on a completely different strategy. For a species to survive and evolve, individual beings have to die and be replaced by fresh ones, and for the life itself to continue, species have to evolve and eventually disappear, replaced by ones more fit to changing conditions.
So let's forget about actual immortality. We have many technical means to record and keep alive for as long as possible the memory of those who are gone, if they deserved it. To our transhumanists I would suggest to simply make their lives something worth remembering. It's a proven recipe for the only kind of immortality which is worth it, the one living in our memories.

[This post is available in French here]

Handwriting questions and answers

Why stick to handwriting? 

It's so painful and slow! 

There are so many efficient technical ways to write and communicate now.

It might be good for the museum and art school, for poetry and diaries. 

But why should I bother?

My handwriting is so ugly anyway. 

Why should I show it at all and why should I make others suffer from deciphering it?

But it shows too much about me ... I don't want it to be analyzed by graphologists.

In other words ...

In praise of handwriting

This started by a post shared by +Teodora Petkova, suggesting to share handwriting, I found the idea was cool, so I started a Google+ collection. For the record here is my today contribution - complete with spelling mistake (thousands of times)

Otherwise said in French

I have started with the new year a kind of mirror of this blog in French, a long overdue return to my native language. I hope some readers of in other words will be fluent enough in French to also make sense and hopefully enjoy those choses autrement dites. The first posts are listed and linked below, with a short abstract.

• Toute chose commence par un trait on Shitao, the unity of painting, calligraphy and poetry in classical Chinese culture, and how the continuum of nature is divided into things by the single brushstroke. 
• Cosmographie en orange et bleu a "just so story" about the separation of heavens and earth as seen and described by the first ontologist in the first days, and what happened to him on the seventh day.
• L'ontologiste sur le rivage des choses on the illusion of so-called ontologies (in the modern sense of the term) thinking they have said what things are, when they only define how things differ from each other.

To be continued ... stay tuned!

Desperately seeking the next scientific revolution

If you still believe in the ambient narrative on the accelerated path of scientific and technological progress, it's time to read You Call this Progress? on the excellent Tom Murphy's blog Do the Math. I'm a bit older than the author, just enough to have seen a few last but not least scientific achievements of the past century happening between my birth and his one. The paper of Watson & Crick on DNA structure was published in Nature a few days after I was born. My childhood time saw the discovery of the cosmic microwave background and general acceptance of the Big Bang theory, experimental confirmation and acceptance of the plate tectonics theory. While I was a student the standard model of microphysics was completed. Meanwhile chaos theory, of which mathematical premices had been discovered by Poincaré at the very beginning of the century, was setting the limits of predictability of natural systems evolution, even under deterministic laws.

This set of discoveries was somehow the bouquet final of a golden age of scientific revolutions which contibuted to our current vision of the world, starting in the 19th century with thermodynamics, theory of species evolution, foundation of microbiology, electromagnetism unification, followed at the beginning of the 20th century by relativity and quantum mechanics, two pillars for our current understanding of microphysics and cosmology, from energy production and nucleosynthesis in stars to structure of galaxies and visible universe at large. Put together, those revolutions spanning about 150 years from 1825 to 1975 set the basis for the mainstream scientific narrative, giving an awesome but broadly consistent (if you don't drill too much in the details, see below) account of our universe history, from Big Bang to galaxies, stars and planets formation and evolution, our small Earth and life at its surface, bacteria, dinosaurs and you and me. A narrative we've come to like and make ours thanks to excellent popularization. We like to be children of the stars, and to wonder, looking at the night sky, if we are the only ones.

As Tom Murphy clearly arguments, this narrative has not substantially changed since 40 years, and has not seriously been challenged by further discoveries. Many details of the story have been clarified, thanks to improved computing power, data acquisition, and spatial exploration. We've discovered thousands of exoplanets as soon as we had the technical ability to detect them, but that did not come as a surprise, and in fact what would have been really disturbing would have been not to discover any. The same lack of surprise happened with gravitational lenses first discovered in 1979 but predicted by general relativity. And no new unexpected particle has been discovered despite billions of dollars dedicated to the Large Hadron Collider, the largest experimental infrastructure ever built.

Could that mean that the golden age of scientific revolutions is really behind us, and all we have to do in the future is to keep on building on top of them an apparently unbound number of technological applications? In other words, that no new radical paradigm shift, similar to the ones of the 1825-1975 period, is likely to happen? Before making such a bold prediction, it would be safe to remember those famous for having proven wrong in the past in pretending that there was nothing new to be discovered.

Actually, major issues already known by 1975 are still open. In physics, the unification of interactions needs to solve strong inconsistencies between relativity and quantum theory, an issue with which Albert Einstein himself struggled until his death, not to speak about the mysterious dark matter and dark energy needed by theory to account for the accelerated expansion of the universe. The latter is actually one of the rare important and unexpected discoveries of the end of the 20th century. In natural science, the process of apparition of life on Earth has still to be clarified, as well as the correlative issue of the existence of extraterrestrial life.

The number of scientists and scientific publications since 1975 has kept growing exponentially, as well as the power of data acquisition, storage and computing technology. With no result comparable in importance for our understanding of the universe to what Galileo discovered in the single year 1610 simply by turning the first telescope towards the Moon, Venus and Jupiter. The general process of science and technology evolution in the past has been that improved technology and instrumentation yields new results pushing towards theoretical revolutions and paradigm shifts. But strangely enough, the unprecented explosion of technologies since half a century has produced nothing of the kind.

Is it really so? Some scientists pretend that there actually is a revolution going on, but as usual mainstream science establishment is rejecting it. This is for example the position of Rupert Sheldrake in this article of 2012 The New Scientific Revolution. Indeed, the theories Sheldrake is defending, such as Morphic Resonance and Morphic Fields, are really disruptive and alluring, but refuted as non-scientific by the majority of his peers. I'm not a biologist, so I won't venture in this debate, and let readers make their own mind about it.

Two cents of (natural) intelligence

Several months ago, my previous attempt to speak here about artificial intelligence, wondering if computers could participate in the invention of language, met a total lack of feedback (it's not too late for second thoughts, dear reader). I found it quite frustrating, hence another attempt to venture on this slippery debate ground.
+Emeka Okoye in the follow-up of the previous post on facets makes strong points. When I wonder how much intelligence we want to delegate to machines, and for which tasks, the answer comes as a clear declaration of intention.

We are not delegating "intelligence" to machines rather we are delegating "tasks" ... We can have a master-slave relationship with machines ... We, humans, must be in control.

I appreciate the cautious quote marks in the above. But can it be that simple? Or just wishful thinking, as +Gideon Rosenblatt is warning us in a post entitled Artificial Intelligence as a Force of Nature. The connected machines ecosystem, distributed agents, neuronal networks and the like, are likely to evolve into systems (call them intelligent or not is a moot point) which might soon escape, or has already escaped if we believe some other experts on this topic, the initial purpose and tasks assigned by their human creators, to explore totally new and unexpected paths. This hypothesis, not completely new, is backened here by a comparison with evolution of life, of which the emergent ambient intelligence would be a natural (in all meanings of the term) follow-up.

But evolution of technologies, from primitive pots, knifes and looms up to our sophisticated information systems, is difficult to compare to the evolution of life and intelligence. The latter is very slow, driven by species selection on time scales of millions of years, spanning thousands of generations. Behind each success we witness, each species we wonder how it perfectly fits its environment, are forgotten zillions of miserable failures which have been eliminated by the pitiless struggle for life. Nothing can support the hypothesis of an original design and intention behind such stories.
It's often said, like in this recent Tech Insider article, that comparing natural and artificial intelligence is like comparing birds to planes. I agree, but this article misses an important argument. Birds can fly, but at no moment did Mother Nature sat down at her engineering desk and decided to design animals able to fly. They just happened to evolve so over millions of years from awkward feathered dinosaurs, jumping and flying better and better and we now have eagles, sterns and falcons. On the contrary, planes were from the beginning designed with the purpose of flying, and in barely half a century they were able to fly higher and quicker than the above natural champions of flight.

To make it short, technology evolves based on purpose and design, life (nature) has neither predefined purpose nor design. Intelligence makes no exception to that. Natural intelligence (ants, dolphins, you and me) is a by-product of evolution, like wings and flight. We were not designed to be intelligent, we just happened to be so as birds happened to fly. But computers were built with a purpose, even if they now behave beyond their original design and purpose, like many other technologies, because the world is complex, open and interconnected.

Let's make a different hypothesis here. Distributed intelligent agents could escape the original purpose and design of their human creators, maybe. But in such a case, they are not likely to emerge as the single super intelligence some hope and others fear. Rather, like the prebiotic soup more than three billions years ago, its spontaneous evolution would probably follow the convoluted and haphazard paths of natural evolution, struggle for survival and the rest. A recipe for success over billions of years, maybe, but not for tomorrow morning.

Rage against the mobile

The conversation around the previous post about facets led me to investigate a bit more about mobile, and what it means for the web of text. This is something I'd never really considered so far, and thanks to +Aaron Bradley for attracting my attention on it. Bear in mind I'm just an old baby-boomer who never adopted mobile devices so far, touchscreens drive me crazy, and I still wonder how people can write anything beyond a two words sentence on such devices etc. To be honest I do have a mobile phone but it is as dumb as can be (see below). It's a nice light, small object, feeling a bit like a pebble in my pocket but I actually barely use it (by today standards), just to quick calls and messages. Most of the time I don't even carry it along with me, let alone check messages, to the despair of my family, friends and former colleagues. But they eventually get used to it.

To make it short, I do not belong to the mobile generation, and my experience of the Web has been from the beginning, is, and is bound to remain a desk activity, even if the desktop has become a laptop along the years. I'm happy with my keyboard and full screen, so why should I change? And when the desk is closed, I'm glad to be offline and unreachable. I wish and hope things can stay that way as long as I'm able to read, think and write.

With such a long disclaimer, what am I untitled to say about mobile? Only quote what others who seem to know better have already written. In this article among others I read about the so-called mobile tipping point, this clear and quite depressing account of the consequences of mobile access on Web content.

The prospect for people who like to read and browse and sample human knowledge, frankly, is of a more precipitous, depressing decline into a black-and-white world without nuance [...] The smaller screens and less nimble navigation on phones lend themselves to consuming directory, video, graphic and podcast content more easily that full sentences. If the text goes much beyond one sentence, it is likely to go unread just because it looks harder to read than the next slice of information on the screen. [...] Visitors who access information via a mobile device don’t stay on sites as long as they do when using a desktop computer. So if you’re counting on people using their smartphones or tablets to take the same deep reading dive into the wonders of your printed or normal Web page messages, you’re probably out of luck.  

Given the frantic efforts of Web content providers to keep audience captive, all is ready for a demagogic vicious circle of simplification. Short sentences, more and more black-and-white so-called facts. If this is where the Web is heading to, count me out. I won't write for mobile more than I use mobile to read and write.

I still have hope, though, looking at this blog analytics. Over 80% of the traffic seems to still come from regular (non mobile) browsers and OS. But I guess many of you visitors have also a mobile (smart) phone you otherwise use. I wonder if and how you manage to balance which device you use for which usage. Are you smart enough to use mobile for apps, and switch to proper desk screens to take the time to read (and write)? I'm curious to know. 

In praise of facets

Follow-up of the previous post, and more on the ways to escape the tyranny of entities in search results. In the quick exchange with Aldo Gangemi in the comments of this post, facets were suggested. I won't argue further with Aldo about facets at BabelNet being types or topics, because he will win at the end, and such a technical argument would lead us astray, far from the main point I wouyld like to make today. You might be uneasy on what facets and particularly faceted search mean, but you have certainly used them many times when searching e-commerce sites, to filter hundreds of laptop models by price, brand, screen size, memory size etc. Libraries, enterprise portals, and many more use faceted search, example below is the search interface of Europeana for "impressionism", the results being filtered by two facets, media type "image" and providing country "Netherlands".

Faceted search is a very intuitive way to search items in a data base. Using faceted search, the user creates at will its own algorithm of filtering, selection and possibly ranking. If you compare with the usual general search engine results, two major advantages appear. The search is multidimensional, and the algorithm is transparent to the user. The system does not apply fancy, smart but opaque algorithms, based on guesses of what the user is looking for. It provides an interface where the user's natural intelligence can be put into action. In short, faceted search provides a good collaborative environment where artificial and human intelligence work together, the former at the service of the latter.

Given the above, one can wonder why general search engines such as Google do not propose faceted search facilities over their results, instead of an unidimensional list of ranked results. A technical answer coming to mind is that such engines do not search items in a collection of objects of which semantic descriptions are stored in a data base, but resources indexed by keywords. That used to be true, but the argument does not seem to hold anymore in the current state of affairs. The Knowledge Graph, however it's implemented, is a data base where things have declared semantic types and properties which could be used for faceted search. It would be a good way to see types and properties defined by schema.org vocabulary put explicitly into action as facets (Creative Work, Person, Place, Event, Business, Intangible ...).

I cannot imagine that Google and al. have never thought about this. There are certainly technical hurdles, but I can't imagine they could not be solved. So I would be curious to hear what they have to say, given that the added value to the search experience would be tremendous. Above all, it would give back to the user the power to define her own filtering on results, and reinstate the habit to do so, instead of the reductionnist Q&A dialogue which in the long run leads to pernicious intellectual laziness, unique thought, and jumping to conclusions without further checking. Our world is more and more complex, and offering simplified and unidimensional answers (presented as facts) to any question does not help to cope with complexity. Current events show us too many examples of oversimplifications and where they lead to.

I think of any of my queries to a search engine as a beam of light sent through the night of my ignorance, where possible answers are hiding as so many complex multi-faceted diamonds. I don't want any one of them, however brilliant and wonderful, make me blind to the point of missing all the rest. Every faceted answer should reflect back a new and unexpected part of the spectrum, without exhausting the question we should always keep alight.

Search is not only for entities

The Knowledge Graph is a great achievement, but its systematic use at the top of search results is sometimes counter-productive. Knowledge Graph nodes are mostly named entities (individuals, particulars) such as people, places, works (movies, books, music tracks), products ... and rarely universals (concepts, topics, common names). And if an ambiguous search sentence can refer to either particular entities or universals, the former seem to always float at the top with their fancy Knowledge Graph display, and relevant results about universals kicked down. The assumption underlying this default behavior is that people search mostly for particular entities (things), not information about some universal (topic). The hijacking of common names as brand names we already pointed here in the past adds to the issue, along with the growing number of work titles using common names. Add to this the magic of the Knowledge Graph knowing entities by various names in different languages, and you end up with examples like the following. 

For a recent post I searched about the Theory of Everything. If instead of going straight to the Wikipedia article I ask Google, here is what I get.

I was searching for information about a theory in physics, and I get all about a movie which happens to have taken as title the name of this theory. And since my browser default language is French, the Knowledge Graph is kind enough to present me the movie under its French adaptation title "Une merveilleuse histoire du temps", which you can imagine even if you don't speak a lot of French, is all but a translation of "Theory of Everything". The silver lining is that if I search for "Théorie du tout" in French, I have not the same problem, since the movie is not known in French under this title which would be the correct translation of the original one. The first result for "Théorie du tout" is the Wikipedia article on this topic, as expected.

You can play the same funny game with "Gravity", "Frenzy" and many more. Given the limited supply of common names, and the exponential growth of named entities in the Knowledge Graph, all tapping into the commons for their names and titles, such ambiguities are likely to end up being rather the rule than exceptions. Search engines should provide a simple way to opt out entities, so that I could ask "Dear Google, give me resources about the topic called gravity, and I don't care about any individual entity with gravity in its name." And yes, Google, you can do it, I'm sure, just take example on BabelNet, where you can sort results by entities, concepts, music, media etc.  A bit of typing goes a long way ...

Why do I write, really?

Teodora Petkova strikes again with her new and tiny (her word) Web Writing Guide. Her savvy recommendations on the Whys and Whats of writing on/for the Web made me wondering if I ever applied any single one of them, and in particular in this blog which has been for years the main place I've been writing and publishing. The rest of my publication track consisting in a handful of conference or journal papers, a chapter in a collaborative book, some of those still published online, but not really written "for the Web". Not to mention hundreds of messages to various community lists and comments on the social Web, but does that really count as Web writing? 

It might be too late and pointless anyway to consider those recommendations, since I have no tangible reason to keep on writing altogether. Retired from business for half a year, not participating any more in discussions of various communities, not even following them, I have nothing to sell or even to give away here. I could as well forever hold my peace, instead of indulging in more wordy selfies. Nevertheless, I'll make the exercise of going through some of Teodora's recommendations, to see if I ever met them. Just for the fun of it.

• Write for people

Of course, who else? But I've never thought of anyone in particular as the target of what I write here, although I know I write better when I think about a potential reader. Somehow, each post on this blog could (should) be read as a personal letter to some unknown reader. To make it short, I have no market, no target audience. I know I have a handful of more or less faithful followers, and hope the few serendipitous visitors will bring home some food for thought. 

• Write for machines

Believe it or not, I really don't give a damn about that one. I've been a so-called Semantic Web evangelist because I liked the ideas behind it and the conceptual debates it triggered (not to mention I was also paid for it), but I never applied its technology to this blog. I even did everything to blur the radar of search engines by changing both URI and title several times. No semantic markup either, beyond a few (rather random) tags. I like the idea of those pages being as easy to reach as the places I love in my mountains. Not unreachable, but not much advertised either, with paths not difficult to follow, but not obvious to find either. And actually, since I'm not able to define or name what I am about, I prefer search engines to ignore those pages than indexing them under any silly topic.

• Write for joy

This is certainly the only recommendation I follow. Nothing to add.

But the Whys are not the main point of difficulty. Regarding the Whats, I must admit I am completely off track.

• What is it that you really want to say and cannot help but share?

I'm afraid most of the time I don't know before I've finished writing it.

• What is it that your audience needs?

As said above, I've no audience, and therefore cannot possibly know what it needs. 

If I try to apply the following ... The intersection of the answers to these questions is the answer to “What to write?” Well I won't say this intersection is empty, but it looks rather undecidable.

Sorry, Teodora, but your recommendations are either useless to me, or they lead to the conclusion that I should not write at all before answering the two above. Unless the write for joy is enough of an excuse to keep writing. 

When I was a child, half a century ago, my school teacher (who happened to be also my father, the teachers offer is scarce in village schools) was an adept of the texte libre. This is the writing exercise I still prefer. Following the Freinet pedagogy, the original free production was selected and amended by the group and eventually published in the class journal. The final text was a collective production based on an individual original idea. Does not that sound quite Webby, back in the 1950's, in remote French village schools?

Backtracking signs

This image has been for some years now my avatar on various places on the Web. I've chosen it obviously because it's a nice image taken in my dear mountains, but also as an illustration of what Quine called the inscrutability of reference.
This image is a sign, elle nous fait signe. To each of you, depending on your experience and culture, it will evoke something different and particular  - or nothing at all. But does it only evoke, or does it represent something? Could a machine figure what it is? I would be curious to submit this image to some automatic description algorithm. Would we get something like tracks in the snow in a winter mountain landscape? That would not be bad. If it succeeds in adding several people wearing snowshoes, I would be most impressed. And I would be really baffled if it could guess how many people have passed, and in which direction.

Now let's take it as a support for an exercise in backtracking. Let's move a few steps towards the genesis of this image, trying to figure out its deeper meaning. Someone shot this image on a fair winter day (supposing it's a genuine photograph and not one of those fancy computer-generated graphics). In either case what you are viewing here and now is just a reconstruction on the screen of your device of a pack of bits, a file uploaded to Google servers from my computer, this local file being itself a resized and trimmed copy of an original one generated by a numeric camera. Several copies, deconstructions and reconstructions happened since the original shot.
Now just trust me it's a "genuine" photograph of some "real" landscape, and imagine yourself back at the scene, along with the photographer. Given the point of view, he's certainly on the tracks himself. Does he follow the tracks let by another group of walkers? Does he belong to this group? Is he looking back at its own tracks? Has he followed the same track way up and down, and the several people who seem to have passed here were actually the same person, once walking up and once down, or maybe several times up and down? Whatever. Who could answer those questions now, except the one who shot the image? Days, months, seasons and years have passed since. Later on the same day other walkers have come following the tracks or crossing them and messing the signs. A few days after a new snow fall has erased them all, and in April the winter memories have vanished in the streams joyfully cascading down. And another summer, and another winter. Going back there now won't tell you anything about those tracks, even if the landscape looks quite the same, even if some walker has taken today the same path, letting similar tracks.

But figuring the genesis of the image itself is not the end of the backtracking. I've chosen this image to represent me on the Web, among thousands of possible images. How can you interpret this choice? Is it a track of mine, captured by someone else, a track of someone else taken by me, my own track taken by myself, a far-fetched form of selfie? Maybe nothing of the sort. Maybe I found this image somewhere on the Web and thought it looked like me, someone who walks, and is often no more where you expected to meet him.

I could answer all those questions, but I won't. I'd rather imagine you wondering as you would wonder, hopefully, finding some perfect pebble stone on the seashore, about the long story it silently tells, the slow cooking of rock in the depth of Earth and its upraising over millions of years, the sudden earthquake or storm or the patient bite of ice cracking the rock, the fall off the cliff, the long rolling travel downstream to the sea, the patient work of currents, tides and waves until this unique morning where its glow on the sand have captured your eyes.

Think about it, just every thing is somehow akin to this image of a track or that pebble stone. Telling stories, giving time its depth by linking us to the past as so many threads. Trees and rocks, bowls, clothes, jewels, printed words and texts. And every so-called Web resource. They are not just sitting idly here and now, but are signs worth backtracking.

Do things go wrong, or is it just me?

Consistency seems to be an universal requirement for any account of reality we accept to consider as true. This requirement seems to build slowly in childhood with the acquisition and consolidation of language, along with notions of true and false, and the underlying law of excluded middle, a basis for all rational and scientific accounts of the world. Formal logic and mathematics underlie the growing computational power of our machines, and we also try to make consistent the laws and rules governing our daily life. But whatever the level of formality at which they are used, consistency and truth belong to the realm of discourse. Holding that a discourse is (in)consistent and statements are true or false in the framework of this discourse makes sense and in many cases can be precisely defined and proven by logic. Considering that a statement is true because it seems consistent with reality or at least the state of affairs at hand is more hazardous, but is still useful and is actually the basis for most of our daily decisions. 

But what is more arguable is to consider consistency as a characteristic of the reality itself, independently of any discourse we can have on it. What could that mean? Reality simply is what it is whether we think or speak about it or not, and there is no point in asking if reality is true or false, consistent or inconsistent, all qualifiers which should apply only to statements and discourse. Reality is the state of affairs, the mountain as we experience it, it is not a discourse, even if our discourse is part of it. What can be said true, false, consistent or inconsistent, is that one asserts about this experience. But somehow the experience has those permanent patterns which comfort us in believing that indeed reality is internally consistent and our language can build accounts of it we proudly call facts. Our faith in the internal logic and consistency of reality beyond any account of it has gone as far as considering reality as the embodiment of the discourse of some perfect logos. This metaphysical stance pervades implicitly or explicitly all the occidental thought from Greek philosophy through various avatars of monotheism. We can still track it in modern science, with the quest of the Theory of Everything, which in the mind of some would be a consistent account of no more no less than the thought of God. Of course such a theory should be globally logically consistent, since the creator could not be inconsistent without failing to perfection.

Our philosophy should be more humble. Logic should stay where it came from and belongs, inside language. And when the reality suddenly behave in an unexpected way, inconsistent with those accounts we so far considered as true, instead of thinking first that things have gone wrong, let us admit that it is our account of things which was proven wrong. Things never go wrong, but we often do.

The moving shores of things

I would like to dedicate this post to the victims of last week's attacks in Paris, who were blindly sentenced to death without notice because they were guilty of joie de vivre, or maybe simply of humanity. I started writing those lines before the attacks, and they could seem at first sight to have nothing to do with Daesh madness. But if you are patient enough to read down to the end, I hope you will find relevant food for thought in the context of those events.

Whether things are ontological primitive or abstracted from the states of affairs, as suggested in our previous post, to deny them any kind of existence would fly in the face of common sense and experience. Mountaineers know that there are mountains, rocks and streams, sailors know that there are seas, waves and storms. But ask them what mountain or sea is, and you're likely to get all but a definition. They will certainly tell you awesome stories of climbing and sailing, maybe show you images, and the most sensible of them will just propose to go with them for climbing or sailing to figure by yourself, experiment the thing, be part of it. 

Why is it so? Because being is not being neatly defined. Our logicians and ontologists would like to make us believe that the world can be split neatly between this and that, day and night, land and sea, human and non-human etc, categories which could be logically defined. There are many reasons why they are wrong, the most often quoted being the arbitrary choice of such limits, since the world can be split into things in many ways. A good introduction to the current discussion on this viewpoint called relativism can be found in the Stanford Encyclopedia of Philosophy.

But relativism is not the best and primary stance I would choose to argue why trying to give a logical definition of mountain or sea or whatever else is bound to fail forever. The main point is that such things, as well as most things you can think of, have fringes, shores, edges, interfaces ... (the name depending on the kind of things you consider) through which they are less separated from than intertwined with each other. If you look at a shore from far enough, it can look like a neat line. But if you look closely, while walking on a beach or at the edge of a forest, you will discover a very complex world which belongs to neither or both worlds that meet here. On the shore, the sea enters the land and the land feeds the sea in the perpetual circulation of waves and tides. The shore is where land and sea communicate and exchange. At the forest edge, animals perpetually come in and out of the trees' shelter to feed in the grass and fields. And what is the forest itself, if not a shore between earth and sky, with thousands of trees as so many links and knots between the depth of ground and the air. Great examples of such intertwining interfaces are mangroves, known to be extraordinary rich ecosystems.

Mangrove at Cayo Levisa, Cuba.

Source Wikimedia Commons

The concepts we abstract from the world and toss to each other's face in our endless arguments and wars are of the same nature. Between life and death, human and non-human, the limits should look indeed like the above, moving and intertwined. And thinking otherwise that those moving shores are or should be reduced to neat lines is the first step towards totalitarism, exclusion, and death. 

The words and acts of Daesh have gone of course very far down such an alley, but to fight them back we should be careful not to use similar simplistic rhetoric, ignoring the complexity of fringes and shores, replacing them with edges as straight and cutting as their knife's blade. One of those, the most simplistic one still unfortunately thought aloud by too many people, leads to broadly confuse Daesh with Islam, when 99.9% of Muslims condemn the terrorism, and more than 80% of Daesh victims are Muslims. But the defensive stance of many moderate Muslims claiming outloud that Daesh is not Islam, and his members are not the Muslims they claim to be, is equally simplistic and counterproductive. Daesh is indeed a shore of Islam, although a very remote and dangerous one, and moderate Muslims would certainly benefit to acknowledge outloud that such a shore exists, where Islam meets and intertwine with intolerance, obscurantism, organized criminality, thirst for glory and power, or sheer madness. As any shore, you can get there from both sides. From inside Islam through fundamentalism, and from outside through social exclusion and criminality. 

Beyond or inside Islam, we see many people saying or writing that Daesh killers have put themselves by their words and acts deliberately beyond humanity, and therefore could and should be simply shot down as dangerous furious animals. Calling them "monsters" or "barbarians", whatever fits for saying "they are not like us and must be eliminated" is as simplistic and counterproductive as the above claim "they are not Muslims". Daesh killers certainly dwell on some strange and frightening fringe of humanity, so far off that they are even able to shake the notions we have of what makes humanity. But whether humans or barely so, barbarians, monsters, or simply mad criminals, in any case they have come to this deadly shores from inside humanity, and we need to understand how they got there to prevent more young people to follow the same paths.

Mountains as states of affairs

Trying to make sense of the deep work of Jan Christoph Westerhoff about ontological categories, reality and everything, along with a slow but steady learning of Chinese language and ancient philosophy, leads you to consider as ontological primitive the states of affairs, instead of good old semantic web things and properties, the latter being derived artefacts of the former, not the other way round. Let's try to illustrate this as simply as possible.

Consider a mountain. On the semantic web you represent a mountain as an instance of owl:Thing or one of its specific subclasses such as schema:Mountain. You claim to have defined a non-ambiguous individual identified by a URI and described by an open set of property-value pairs, such as http://dbpedia.org/resource/Mount_Everest.

But in the view of the world proposed by both Westerhoff philosophy and the Chinese language (insofar as I understand them properly), the above are just abstractions derived from some state of affairs. The chinese 山(shān) we translate in English as mountain(s) is a sign associated with certain aspects of things, or states of the world. We have to be very cautious on terms here, and not take for granted that existence of "things" and "the world" are preconditions to the states of affairs we associate with the sign 山. In ancient Chinese culture where this sign first emerged about three thousands years ago, the world is not divided into things before we name them. Certain states of affairs, patterns we recognize again and again, lead us to associate a sign to them. 山 is just an abstract visual representation of those states of affairs presenting peaks rising upward, a main central one and another one of each side, slightly asymmetrical. A mountain is indeed generally mountains, bearing in mind "three" has to be understood as a shortcut for "many".

The difference between considering there are such things in the world as individual mountains and we just give them individual names and put them in a category, and considering mountains as states of affairs we associate using a common sign or name, might appear subtle or moot. But it is indeed a fundamental shift of our view of the world. States of affairs are not neat individuals defined by properties, they are not separated from each other, they have neither precise limits in space and time, nor definite components and properties. Of course we can try to agree and generally agree to disagree upon such limits and components, and argue forever on what is or is not a mountain in general or this mountain in particular. And we actually argue upon what is a human being, or a book, or a Web resource, or democracy ... This kind of argument is interestingly called in Chinese 是非 (shì fēi), literally meaning "being - not being", hence "right - wrong" and in common language dispute, argument. There is much food for thought in this word. Dispute arises when the language gets out of its original role of simply putting signs on state of affairs, going down to argue on what there is and is not behind signs, in other words, when the language mingles into ontology and meaning instead of sticking to what it's really made for - poetry.

I wish you to stay away from dispute, walk up and listen to the mountain songs.