Jaron Lanier on Antifragile Software

March 23, 2013 § Leave a comment

Aren’t bugs just a limitation of human minds?

No, no, they’re not. What’s the difference between a bug and a variation or an imperfection? If you think about it, if you make a small change to a program, it can result in an enormous change in what the program does. If nature worked that way, the universe would crash all the time. Certainly there wouldn’t be any evolution or life. There’s something about the way complexity builds up in nature so that if you have a small change, it results in sufficiently small results; it’s possible to have incremental evolution. Right now, we have a little bit — not total — but a little bit of linearity in the connection between genotype and phenotype, if you want to speak in those terms. But in software, there’s a chaotic relationship between the source code (the “genotype”) and the observed effects of programs — what you might call the “phenotype” of a program.

Right. And it results in a type of error that doesn’t teach you anything. You have chaotic errors where all you can say is, “Boy, this was really screwed up, and I guess I need to go in and go through the whole thing and fix it.” You don’t have errors that are proportionate to the source of the error. And that means you can never have any sense of gradual evolution or approximate systems. So, the real difference between the current idea of software, which is protocol adherence, and the idea I’m discussing, pattern recognition, has to do with the kinds of errors we’re creating. We need a system in which errors are more often proportional to the source of the error (emphasis added).

So pattern recognition is really starting to come into its own. Sadly, a lot of that’s driven by security and defense requirements, but for whatever reason, it’s becoming viable. And we’re at the point where computers can recognize similarities instead of perfect identities, which is essentially what pattern recognition is about. If we can move from perfection to similarity, then we can start to reexamine the way we build software. So instead of requiring protocol adherence in which each component has to be perfectly matched to other components down to the bit, we can begin to have similarity. Then a form of very graceful error tolerance, with a predictable overhead, becomes possible. The big bet I want to make as a computer scientist is that that’s the secret missing ingredient that we need to create a new kind of software.


Research project: Make software robust or antifragile so that small changes do not blow up.


There Is No Such Thing as Bar School

March 17, 2013 § Leave a comment

Bar school is the mandatory four or eight month professional legal training program organized by the Quebec bar. Students attend bar school after graduating from law school and before doing their articles. It consists of, among other things, four hours of class five days a week, two midterm exams, and a two-day final exam. As there are a number of requirements (e.g., assignments, conferences, and training sessions) that are not tested on any of the exams, bar school is not only about exam prep but about developing a set of core competencies required for the practice of law. As far as I know, this kind of practical training program is unique to Quebec, at least in North America. Bars in common law jurisdictions, both in Canada and the United States, do not provide any practical training, relying on the law schools to perform that function. To get a licence to practice law, you merely need to prove your competence by passing the bar exam. There are of course prep courses, but these are not the same thing as bar school because unlike bar school their sole purpose is to prepare students for writing the exam.

Surprisingly, there is no such thing as bar school.

The meaning of a noun is the thing it refers to, as determined by its sense. The sense of a noun is a description or concept, that is a set of necessary and sufficient conditions that picks out one or more things in the world. For instance, the sense of “Glenn Gould” might be “the greatest Canadian pianist”. This sense would pick out a particular person, namely Glenn Gould, which is the meaning of the words “Glenn Gould”. The meaning of a noun is the meaning attributed to it by a community of competent speakers. Just because one person assigns some other meaning to a noun does not mean that that is the meaning of the noun. For example, if one person thinks that the sense of “Glenn Gould” is “the greatest Prime Minister of Canada”, that does not mean that “Glenn Gould” refers to the greatest Prime Minister of Canada, whoever that might be.

There is no such thing as “x” could mean that the sense of “x” does not pick out anything in the world. There is no particular thing that “x” refers to. Or, in other words, “x” refers to the empty set. Examples of this would be “the king of France” or “Montreal’s rugby team”. These compound nouns make sense but they do not refer to anything. This is what you would mean if you were to say “there is no such thing as ‘the king of France’.”

There is no such thing as “x” could also mean that although “x” refers to something, we (i.e., the community of competent speakers) do not use “x” to refer to it. For example, someone could say, “There is no such thing as mortgages in Quebec. We call them hypothecs.” The sense of “mortgage” might be “a loan secured by your house”. And since we do have loans secured by houses in Quebec, “mortgage” does have a referent. It is just that we do not use “mortgage” to refer to that legal practice. We use the word “hypothec” instead. If someone in Quebec were to say “I just got a mortgage on my house”, it would be legitimate, although maybe a little pedantic, to say “No you didn’t, there is no such thing as mortgages in Quebec. You must mean that you got a hypothec.”

There is no such thing as “bar school”. Having spoken with a number of members of the community of competent English speakers, this community does not know what bar school is. If anything, they tend to associate it with the prep schools of common law jurisdictions. This means that for the typical competent English speaker, the sense of “bar school” is “a prep school for taking bar exams.” Since there is no prep school in Quebec for taking bar exams, “bar school” refers to nothing. To be sure, “bar school” does refer to the correct institution in the legal community. There is nonetheless no such thing as “bar school”, in the first sense described above, among the broader community of competent English speakers.

Yes, I hear you say, but there are prep schools for bar exams out there in the common law jurisdictions. Wouldn’t “bar school” then refer to these? It certainly does, but as far as I know, common law jurisdictions do not use the words “bar school” to refer to their prep schools. I can therefore legitimately say that there is no such thing as “bar school” in the second sense described above.

If there is going to be a problem with this argument, I suspect it will be at the empirical level. I have not done an extensive study of the community of competent English speakers. Perhaps “bar school” really does refer to the mandatory professional training program put on by the bar of Quebec. Or, perhaps there is some common law jurisdiction out there that uses the words “bar school” to refer to its prep courses. In either of these cases I will be wrong that there is no such thing as bar school. It would still nonetheless be interesting that starting from pretty orthodox assumptions about the philosophy of language and certain empirical facts, it is possible to reach the implausible conclusion that there is no such thing as an institution that clearly does exist.

The Effect of Technology on Attention

March 11, 2013 § Leave a comment


  • Technology (especially gadgets like smart phones) makes us less able to focus our attention on a single task and leads to ineffective multitasking.
  • Make use of gaps in the day by allowing your mind to wander. Don’t immediately check your phone or email.
  • We should take gadget holidays (e.g., one day without gadgets per week)
  • We should practice focusing our attention daily (e.g., by taking a 15 minute walk, meditating, religious practice).
  • I would add the following as ways of practicing attention: working on or solving a difficult math, computer science, or philosophy problem.

In Defence of the Professoriate

February 26, 2013 § Leave a comment

The funding of post-secondary education is a hot topic in Quebec. One of the suggestions I sometimes hear is that universities should pay professors less. We should, in other words, make the professors pay for the budget cuts. Professors are certainly an easy target given their high level of compensation and low level of professional obligations. We should, however, not forget the important social good that professors contribute and how cutting professor compensation could harm this social good.

Professors seem to be well compensated. Universities pay their tenured professors a good salary (at McGill for example the median salary for an associate professor is $110,633) for modest teaching duties (I believe a 3/3 course load is common at research universities, which translates to nine hours per week lecturing plus tutorials, preparation time, marking, and office hours) and research, for which professors are unsupervised and free to pursue their interests. Not to mention that professors have summers off and sabbaticals every six years.

If professor compensation goes down, I doubt that tenured faculty would be affected. One of trends over the past ten to twenty years that most people outside of academia are probably unaware of is the rise of adjunct faculty:

Although college costs are out of control, the money isn’t going to the professors. There are 1.5 million university professors in the US and , 1 million of whom are adjuncts. One million professors in America are hired on short-term contracts, most often for one semester at a time, with no job security whatsoever. And earning, on average, $20K a year gross, with no benefits or healthcare, no unemployment insurance when they are out of work (http://goo.gl/verfN).

It has gotten so bad that some professors are now recommending that even their good students not pursue the PhD (http://goo.gl/U3g8a). The numbers might not be as bad in Quebec as they are in the United States. It would nonetheless be fair to assume that if universities cut compensation to professors, the cuts will come not in the form of reduced pay to already tenured faculty but by replacing tenured positions with adjunct positions when older faculty retire.

This paints a bleak picture for young academics, but why should Quebec society care? We should care because professors produce something of great social value: research. This is not to say that professors should be forced to pursue only socially valuable research. It is very difficult to predict in advance what research will be socially valuable. Research has the most social value when it results from intelligent people pursuing their interests rather than when research programs are dictated top-down to meet some social goal. I also do not want to say that the work of every professor will be socially valuable. Research  has fat tails. Most research will not have social value. But a few results will have huge payoffs for society that will more than justify the investment.

The phenomenon of adjunct faculty has two important negative consequences for society. In the first place, it is a waste of human capital. Adjuncts rarely have the time or resources to pursue their research. Yet if the professor has gotten his PhD at a publicly funded university, we as a society have already invested a great deal in his education. It is a waste to not make use of that education to produce research. We also risk not attracting the best people to academia in the first place. From the point of view of someone considering an academic career today, the salary of tenured professors is not the main concern. The main concern is the difficulty of finding a decent job after graduation. There is no doubt that many bright people are already not going into academia because of the uncertainty. If there is a significant risk of not getting a tenure track job, many of the best people will pursue careers in other fields. If we cannot attract the best people into academia, if being a professor is not a real option for the brightest students, the quality of research will go down and we will lose the value of having people dedicating their lives to research.

Quebec has some difficult decisions to make about who should pay for its post-secondary education system. Before making the professors pay, though, we should remember the long run costs of doing so.

The Nature of Law and Authority

February 22, 2013 § Leave a comment

Usually the nature of law debate is framed as an ontological debate. It is supposed to be about what the law is. But lately I’ve been wondering if the different positions philosophers take in the nature of law debate stem not from different intuitions about what the law is but about authority.

Positivists say that the law is those rules that emanate from the proper sources,  such as the legislature or a judge. For positivism, it is agents of the state who have authority. They have the authority to alter the legal relations between citizens, and they have this authority because it has been transfered to them.

For natural law theorists, the law is authoritative because it coincides with moral rules that are binding on people anyway. The authority of the law derives from the authority of morality.

For Dworkin, the authority of the law stems from tradition. Admittedly, Dworkin is not often read in this way. But on Dworkin’s theory the law is the best interpretation of the existing legal materials and institutions. This means that the authority of the law derives from the tradition of law, which contains those legal materials and institutions.

Each of these positions has a paradigmatic kind of law. For positivists, it’s is statutory law. For natural law theorists, it is certain aspects of criminal law, constitutional law, and  procedure (e.g., audi alteram partem). And for Dworkin, it’s the common law or the civil law (even with codification, though this is a more complicated story to tell).

Some of these positions also have corresponding advocates in Roman law. Natural law theorists are represented by Gaius and Ulpian, at the beginning of the Corpus Iuris Civilis. And Dworkin’s tradition based view is that of the Roman jurists more generally. For the jurists, the source of authority for Roman law stemmed from the Twelve Tables, and the Twelve Tables themselves were merely a writing down of what had always been done. So the authority of Roman law was tradition. If the jurists wanted to adapt the law, they had to frame it as an interpretation of the legal materials they had at hand because that was where they got their authority.

There doesn’t seem to me to be an analogue of positive law in Roman law. Praetorian law doesn’t count, because even the Praetor had to frame his modifications to the law as interpretations, so even his authority stemmed from tradition. Positive law is really a product of the development of states, codification, and the modern predominance of legislative law over common and civil law. Positivism is the most modern theory of law and authority.

I don’t think it even makes sense to think of the authority to make law within the conceptual framework of the Roman jurists. For that, you need the idea of rights as powers that can be transferred from citizen to state, which Roman jurists did not have. More on that another time.

Reply to David Cole: Free Speech and the Harm Principle

January 5, 2013 § Leave a comment

The anti-Islamic YouTube video that lead to rioting in Egypt and Libya in September 2012 demonstrates the ability of speech to cause violence and harm. In a NYRblog post, the law professor David Cole argues that there should be no legal restrictions on hate speech such as this video, even when that speech causes harm. I cannot see why speech acts, unlike all other acts, cannot be prohibited when they cause harm to others.

No doubt, freedom of expression is an important right deserving of constitutional protection. Free speech is a necessary condition for science, the discovery of truth more generally, the arts, and culture. For democracies to flourish, citizens need to have the right to freely criticize those in power, and they also need to be able to express themselves so that their democratic representatives can know their views. Moreover, individual flourishing and self-fulfillment depends on people being able to freely give expression to their deeply held thoughts and beliefs. Freedom of expression is an important right requiring a heightened level of protection. But does it deserve absolute protection? In short, no.

We limit even the most important rights because no right gives its holder the ability to harm someone else. This is called the harm principle, and John Stuart Mill expressed it in his book On Liberty, where he says that “the only purpose for which power can be rightfully exercised over any member of a civilized community, against his will, is to prevent harm to others” (17). Admittedly, the term “harm” is vague. Does it involve only physical harm? Or does it also include emotional or psychological harm? What about monetary losses? Let us leave these issues aside for the moment and deal exclusively with the case where, as with the anti-Islamic YouTube video, the speech in question causes physical harm and results in real violence.

Since the harm principle applies to all rights I cannot see any reason why it should not also apply to freedom of expression. Mill agrees: “An opinion that corn-dealers are starvers of the poor … may justly incur punishment when delivered orally to an excited mob assembled before the house of a corn-dealer” (104). Inciting a mob to violence violates the harm principle for the simple reason that it is an action that causes harm, and therefore no right to free speech can justify permitting it. Justice Holmes expressed a similar view in the well-known passage from Schenck v. United States:

The most stringent protection of free speech would not protect a man in falsely shouting fire in a theatre and causing a panic. It does not even protect a man from an injunction against uttering words that may have all the effect of force. The question in every case is whether the words used are used in such circumstances and are of such a nature as to create a clear and present danger that they will bring about the substantive evils that Congress has a right to prevent.

This passage recognizes the truth that words can have “all the effect of force”, and when they do we need to prohibit them like we prohibit all unilateral acts of force.

Cole bases his concern about prohibiting harmful speech on a kind of slippery slope argument. He says, “[D]efining ‘hate speech’ in a way that draws a clear and enforceable line between that which deserves protection and that which can be prohibited is an elusive, and probably impossible, task.” Cole is concerned that judges and juries will go too far, restrict too much, and that people will self-censor themselves out of fear of punishment under a vague legal definition. This may be a valid concern for those who attempt to prohibit speech that causes offence or harms dignity or equality. The harm principle, however, is far more modest in the speech it prohibits. It only restricts speech that a judge can causally connect to some harm. Most hate speech does not cause harm. It offends, it hurts dignity or equality, but it more often than not does not result in physical harm. Since the harm principle in practice prohibits so little speech, courts are unlikely to take the restrictions too far and prohibit too much speech, and people are unlikely to censor themselves. Cole’s line-drawing concern does not apply to the harm principle.

Cole’s solution to hate speech is a familiar one: more speech. To Cole’s way of thinking, the best way to prevent the harms of hate speech is to allow the victims the right to fight back with more speech. But once you recognize, as Cole does, that speech can sometimes have “all the effect of force” Cole’s solution to hate speech is nothing more than a law of retaliation, which allows the stronger to get away with harming the weaker. Cole does not provide any justification for not protecting against unilateral acts of force merely because those acts of force come in the form of speech.

July 29, 2012 § Leave a comment

A Philosophy Bites interview with Philip Pettit on republicanism.