Buckingham Inquirer

Martin Rhonheimer's Perspective of the Acting Person

2012-02-28T19:05:00.001-08:00

At the 2011 Lonergan on the Edge conference at Marquette University, Gilles Mongeau made the suggestion that Martin Rhonheimer had in The Perspective of the Acting Person, achieved a positional ethics (i.e., one compatible with correct epistemology). In short, he had taken the subjectivity of the human person (critical existentialism) seriously in his understanding of what Thomas Aquinas meant by practical reason, and had nonetheless arrived at a Thomistic understanding of natural law (i.e. one that affirms with Veritatis Splendor that some acts are intrinsically evil). Naturally I was interested and managed to pick up a copy in due course. Having read it breathlessly over four days in which I was supposed to be doing something else, I can say that I think it mostly succeeds in this ambition, which is wonderful news for Catholic philosophy and theology and the church at large. In fact, the last time I was this excited about a work of philosophy I was reading Alasdair MacIntyre's Three Rival Methods of Moral Enquiry, which is good company indeed. I'll be leading an afternoon seminar on the book at this summer's Lonergan Workshop (Boston College, 18-21 June) but I wanted to put some preliminary thoughts to press before then.

Rhonheimer's basic move is to understand the "object" of a moral act as its "basic intention"--the "for sake of which" the act is fundamentally done. This grounds his ethics in cognition, thereby making it positional rather than dogmatic. The question is whether he can give an account of authenticity in subjective intention sufficient to ground objectivity about the intentions of various kinds of acts, which can then be used as leverage to delimit some of them as intrinsically evil. After reading Perspective of the Acting Person twice, and also skimming Natural Law and Practical Reason (I've yet to read Ethics of Procreation and the Defense of Human Life) I'm convinced that he not only has the right premises and conclusions but is also on the right track in resting much of his argument on the phenomenology (well, I am a Lonerganian). The trouble is that I'm not sure he actually does enough phenomenological work to ground his claims. How, exactly, do we know what "for the sake of" is basic? This is the sort of thing that leads his critics to assume that he doesn't offer an objective ethics, though I think it's rather rash to move from "this argument is incomplete" to "this man is a heretic." Of course, it's also the kind of problem that leads many Lonerganians to unfortunately question Veritatis Splendor, so I have a strongly vested interest in the success of this project.

My hope, which I hope to flesh out this summer at Boston College, is that Pat Byrne has already done some of this mental close reading in his forthcoming book on applied ethics, which can be fruitfully paired with Rhonheimer's historical work on Thomas Aquinas and his extensive argumentation in particular fields of application. Obviously this is parochial, but I would love to know if Rhonheimer himself has been exposed to Lonergan given the lack of citations. It seems like they're very much in the same tradition of reading Thomas Aquinas. Do Opus Dei and the Jesuits just not cross paths? I also wonder whether, given his voice of authority on its interpretation, Rhonheimer had a hand in writing Veritatis Splendor itself?

I also have two fundamental questions about his use of the word "justice." He understands justice as a virtue, but one that not only causes acts as a habit, but also enters into the constitution of their objects. This is how he solves the Le Mis problem: Jean Valjean is not stealing from the Bishop (which would be an intrinsic evil) because Valjean's lack is so unjust that the Bishop then has no entitlement to the property, and its removal is not theft. While I think this has to be the right sort of move, it doesn't seem to me like Rhonheimer gives a sufficient account of what prevents this from reducing to circularity. As with the question of "basic intention" he needs to give a phenomenological account of the virtue of justice with an inbuilt criterion of authenticity. Second, why does justice as he has generically defined it allow us to privilege those close to us--in other words what separates him from Peter Singer? Obviously this wouldn't devolve into Singer's broader problems due to the constraints of the overall ethical framework, and an extraordinary duty to remote others may even fit under Rhonheimer's dictum that "Christian Hope makes the reasonable ideal practical" but it seems like he does resist Singer's construal of justice. Since he doesn't specify how or why, I wonder whether this is on Mill-ian grounds (we are better able to understand and effect the good of those near us) or if he has a more sophisticated take like that of Pat Byrne.

Beyond those general concerns I also have questions about some of the particular applied arguments:

1. In his discussion of inebriation on p.86, he distinguishes it from anesthesia by its "basic intention": but is the basic intention of inebriation often not the alleviation of some kind of psychic anguish? Is there a moral brightline between physical and psychic anguish? Where do drugs that relieve psychic anguish of diagnosed disorders fit in? Or if inebriation for this purpose is morally allowed, do we fall back to a golden-mean type test of virtue?

2. I am particularly compelled by his just war argumentation (p.88), but I wonder why it's not possible that the soldier would be simultaneously choosing the death and life of his aggressor while he pulls the trigger. Clearly acts can have two basic intentions (cf. the growing in love of the couple and the procreation of children in the marital act), so it seems like the barrier must be phenomenological. I certainly haven't thought the issue through, but given the frequent reports of being "of two minds," "acting under instinct," "having conflicting feelings," even inexplicably firing in some situations but not others, etc, it seems like some compelling concrete account is necessary in order to reach the conclusion given. Furthermore, where does this take us on Mutually Assured Destruction nuclear policy? Also, how does it relate to just war principles about injury to non-combatants? If the death of the combatant is also undesired though it consequentially follows, what grants higher status to noncombatants?

3. On p.84 he describes masturbation as being for pleasure rather than for the "release of psychic pressure" but is this obviously true? Again, given the confusing and muddled actual accounts, more work needs to be done here.

4. In his p.92 discussion of capital punishment, why is "not wanting the condemned to die" uniquely the "means to the good of punishment"? What framework for the good of punishment is being assumed here?

5. On p.154, what does it mean for acts to be "of generative kind"?

That's all for now folks, but I badly want this project to succeed, so I hope I'll have more to say as I read Ethics of Procreation and the Defense of Human Life and think about how Rhonheimer and Byrne's work interact in preparation for the workshop this summer.

*Edit: Daniel De Haan was kind enough to point me to Stephen Jensen's article in this issue of the American Catholic Philosophical Quarterly which makes basically all of the above points. In short, he's really worried that there's no definitive account of basic intention, which opens everything to circularity. I obviously share that worry, but Lonergan's example gives me hope that a phenomenological criterion of authenticity can turn an apparent circularity (my initial reaction to Insight) into a heuristic structure. The question is where to find the former, and I continue to hope that Pat Byrne's work holds the answer.

In Praise of Fewer Resolutions

2012-01-22T17:36:00.000-08:00

Because Jim Menick is wrong on the internet, I feel compelled to respond. Jim seems to think that we should go back to prepping 5-7 resolutions a year in LD, because prepping resolutions has value and it makes bad resolutions go by more quickly. I'll answer on the line by line and then offer a counterplan with net benefits.

1a: If practicing resolutional prep is a good thing, then it can better be achieved by a single tournament of parliamentary/legislative debate or extemp than a whole season of old-school LD or even modern PF, so the status quo solves.

1b: There's less of a discontinuity between prepping new arguments on a resolution and prepping a new resolution than Jim thinks. Killing one to save many/vigilantes/civil disobedience/response to domestic abuse all have deeply overlapping ground--often moreso than various cases on the same topic, which are written precisely to avoid common objections and thus need to trace out new approaches. This might even be a turn since there's a time tradeoff between adapting old arguments to new resolutions (with observations and link evidence) and understanding new philosophical approaches (e.g. virtue ethics).

1c: Frequently changing resolutions advantage those who can perfect truly generic ground like kritiks or micropolitics. They don't experience the disadvantage of new resolutions, and those who actually prepare for the new resolutions don't have enough time to write good answers to the generics.

1d: Frequently changing resolutions was an adaptation to a world where absent a summer camp in a university library, finding new angles on an existing resolution was very difficult (note how evidence poor LD cases were in the nineties). The internet solves this.

2a: Bad resolutions might go by more quickly, but if those resolutions are the chosen ones for States/TOC/NFL/CFL, they have just as much impact.

2b: Resolutional quality isn't endogenous to resolutional quantity. The wording committee only has so much time to work, and proposal quality would also go up if there were fewer proposals needed (see: policy). This turns the argument.

2c: More resolutions advantage large squads and those with large budgets. Small squads with poor connections to the alumni coaching network need to attend a circuit tournament before they even know what is going to be run on a resolution, and with the resolution changing every two months that means they need to attend one every month. For resolutions that are basically only run once, such squads get completely blindsided by those who can do a lot of brainstorming and scrimmages.

Counterplan: LD should have two resolutions a year, one released August fifteenth and the other December first. Resolutions should be drawn from a very long list which has five added to it every June by the topic committee. Advantages:

1. The list can grow very long (a hundred?), thus making current prep on the entire list infeasible for even very large teams, and lowering the edge provided by going to a summer camp which luckily picked an important topic (and the resentment from paying a lot of money to a camp which failed to do so). This also means that resolutions will tend to spend a long time on the list before being picked, allowing campaigns against those with poor or offensive wording.

1b. This basically sets LD up as a logistically constrained version of policy, which has all the advantages of policy debate theory, but without the semi-mandatory camp and evidence burden and requirement for a partner. As I've argued previously, affirmative parametrics will solve for negative win skew. This is analogous to off-season-practice rules for sports, but without the enforcement hassles.

2. Gets rid of the most important driver for modest novice (that novices can't be expected to debate a new resolution on their second or third tournament) without triggering the harms of modest novice (small squads and those on the edge of the service area are burdened with prepping yet another resolution).

3. Debate quality improves for the Glenbrooks, Blake, Apple Valley, and Princeton just as it has for TOC vs NFL and CFL. The quality of debate has gone up tremendously since the nineties, except for those tournaments which insist on vanity resolutions and random judging. This is not random.

Cluster Filesystems: Some people still don't get HA

2012-01-20T13:34:00.000-08:00

HA is about, more than anything, transparency and defined behavior. I don't lose any sleep over switches, loadbalancers, NetApps, or DRBD systems that take 30s to come up after failover. An issue where I would need to go physically swap a piece of hardware, or manually intervene to fail over between datacenters, might cause an hour of downtime, but at least it's well-bounded. The things that really kill you are the ones where you shoot yourself or lose data. Restoring backups on fresh hardware might take long enough that most people wouldn't describe it as HA, but it takes a lot less time than manual data recovery and change reconciliation.

Wide area block devices aren't transparent and they don't have defined behavior. That makes them a much worse choice than distributed databases, and a bit of setup complexity on the sysadmin and application sides doesn't change that. I like Jonathan Ellis, but I think he just doesn't get it here. He's focused on the technical possibilities at a particular layer of the stack (you can build a distributed filesystem on Cassandra, after all), rather than the performance of the stack as a whole (the only reason you'd do that is if you had an app that only understood filesystems, and then your app can't make the decisions it needs to make given a distributed backend). Developers who think that they can get HA resilience for free on top of someone else's abstractions are dangerously kidding themselves and need to stop. Build your app within the constraints of someone's PaaS, build your own CAS tradeoff logic, or tell the business they can't afford HA. Welcome to reality, folks.

Glad Tidings

2012-01-20T12:29:00.000-08:00

Just a quick post which I'd hoped to have out for Christmas highlighting some of the best recent software releases from a sysadmin perspective.

First, unit testing tools for Puppet. Turning your infrastructure into code is an amazing development for about 50,000,000 reasons, but you'll only fully capitalize if you're willing to learn from the best practices of software development, like version control, peer review, modularization, and testing. Up to now, testing tools for Puppet had been rather clunky, but this looks like a significant improvement.

Second, PAM authentication for MySQL. I'm definitely a fan of Postgres, Cassandra, or HBase over MySQL, depending on the application requirements, but if you're stuck with a legacy MySQL deployment this is a security godsend. I'd keep my application passwords in the legacy system for performance and robustness, but elevated privileges should be strictly the domain of human beings, and this allows you to easily manage those accounts with whatever LDAP or Puppet tooling you've already built for system accounts. Welcome to access rights on day one and clean shutoff for departures. You might think that my MySQL pick would be Percona's synchronous replication, but I have serious performance and robustness misgivings about that arrangement: it's more complicated than simple failover, but doesn't give you the performance advantages of sharding.

Third, Netflix released their zookeeper library. Multi-tier applications need a way to keep track of what cluster members are live in each tier, and traditional solutions like Puppet and DNS are too slow and asynchronous for large deployments. Zookeeper is perfectly built for this problem, but adoption was limited since the interface was a pain unless you had a homogenous java stack. Now solved :).

Fourth, a set of tools for making java deployments easier. I guess Java and the JVM platform are great for developers, with lots of tooling support for rapid development but a reasonable level of access to networking, data structures, and algorithms for performance. It also allows hacked together Mac and Windows development environments. For sysadmins, though, Java has basically been a nightmare, with its supposed portability meaning that it doesn't fit in cleanly with native Linux tools for deployment, configuration, and monitoring. The last may still be a major issue, but the first seems to have made a major step forward with the projects mentioned above.

Fifth, Amazon DynamoDB. No, it doesn't have the features large users of Cassandra and HBase, or even clustered/sharded SQL, have come to expect. It's probably not much cheaper, either. But for many good and bad reasons, lots of OLTP webapps are committed to running in the cloud, and for those (read: nearly all) where SimpleDB wasn't enough, this is a massive improvement over the disaster of running your own SQL or NoSQL databases on block storage.

Happy 2012, sysadmins!

Ballot Signals

2012-01-20T12:08:00.000-08:00

A debate ballot signals a number of different things to a number of different people:

1. Below the line, it tells debaters and coaches why the judge awarded the win, and (implicitly or explicitly) what both debaters can do to improve their odds of winning on that resolution and/or in front of that judge in future. This also serves the secondary effect of demonstrating that other judges would have voted similarly, so that the losing debater continues to prefer the judge. This is of course why we use the signal "tabula rasa" and fight about what it means: not that we think we can truly be mental blank slates, but that we can empty enough of our minds to be content-(rather than skill at listening and reasoning)interchangeable with other judges.

2. Above the line, it tells tournament administrators how to pair future rounds and allocate trophies. Part of this is done simply by the allocation of wins and losses, but logistic constraints mean that unlike in college policy, where there are usually enough prelims and outrounds to determine breaks on W/L alone and other ballot features seem to be aimed at rewarding the stronger member of unequal teams, HS debate uses further subjective features of the ballot to increase discrimination of debaters with fewer rounds. In the absence of teams and associated ranks, this usually means speaker points, though this year's Lexington Round Robin introduced a smiley system.

For the remainder of this post I'd like to look at what ballot signals beyond W/L are appropriate for 1v1 debate, and how they might be structured. There are two problems with the traditional "speaker points" system. One is that on a 20-30 (by wholes, halves, or tenths) system, there is no real agreement on where the curve should lie; the other is that most judges attempt to use speaker points to convey a number of (sometimes?) orthogonal dimensions simultaneously, with predictably confusing effects and subjective weighing among them. I'll address those problems in order.

The first problem, as Menick points out, is that human beings can't reliably differentiate on a memory-bound basis more than seven levels of anything (e.g. color) let alone something as complex as debaters or debates. So any system should be limited to something like out-of-range(low):1..6:out-of-range(high). If there are more choices than that, judges are forced to either use only a portion of the available scale (and of course they won't all use the same one, with some favoring the top, others the middle or bottom, leading to tremendous random variance in points) or will choose randomly within certain bands. The present frequently multi-modal distribution of speaker points bears out this problem.

One might wonder whether, as long as each judge is consistent, debate could use z-scores as International Science Fair does. The problem with this attempt at statistical normalization is that it assumes both that judges are equally likely to see good and poor debaters in any given round and that scores should be normally distributed in the population. Both of those assumptions, however, are empirically suspect. First, any system of judge ranking means that some judges will be more likely to see teams of a certain caliber (on the bubble, etc). Even more trivially, with power pairing, the two debaters in the round are likely to be of a similar caliber. Second, debater quality is not well distributed among schools (e.g. per-team caps probably reduce rather than increase pool quality). Thus the existence of per-team caps, which improve perception of tournament quality (because presumably coaches already know who their own best debaters are, and what we want the tournament to figure out is the inter-scholastic question), would tend to level the quality curve if the quality curve in the natural population is already normal (and if we can't assume that it's normal, then how are we justified in assuming much at all?). Thus it's a bad idea either to normalize judges' scores for them or to encourage them to normalize their own score assignments.

That statistical cloud, however, does have a silver lining: the original justification for ballot signals beyond W/L is that with a somewhat homogenous and long-lived pool of judges, the adjudicators have epistemic access to facts about relative debater position on a broader scale than the individual round, unlike in psych lab color perception studies. Given things like MJP and the reliable (fairly small) pool of "circuit judges" who account for an enormous percentage of rounds judged, this assumption is fairly plausible. That means that instinctive senses like "is this debater in the top 25% of the field at this tournament" are ones a judge might reasonably make based on past experience with that debater and that tournament. Menick thus labels my 0-7 scale concretely from "wins the tournament" to "needs work to compete in this division" and I think this model is very statistically sound. The obvious problem with this model is that, like any attempt to get the judge to correlate the debater with a broader frame of reference, race/class/gender discrimination issues become more likely, since the judge is working with a more intuitive paradigm. This may be a fundamental tradeoff with more rounds which would require further statistical analysis to quantify.

The second problem is that judges are using speaker points to cover measures as uncorrelated as politeness, rhetorical ability, fit with the judge's aesthetic preferences for debate, tactical prowess, and closeness of round. My proposal would be to allot these as separate scales, with 0(needs work to compete in division)-7(should win tournament) for tactical prowess, 0(offensive)-7(deeply satisfying) for aesthetic quality (smileyness), and 0(tournament director flipped a coin)-7(absolute blowout) for closeness of round. The first factor would be used to pair debaters in future prelims, the second to give non-placement awards (speaker awards), and the third, along with aff/neg split could be used to weight opponent wins for deciding who breaks. Obviously these would all require tabulation software support to be practicable at larger tournaments, but I don't see any reason for conflating these plausibly uncorrelated factors (empirical support: lack of correlation between smilies and placement at LexRR; emotional support: deep angst of judges about what speaker points to allocate to debaters who personally or aesthetically offend them).

New Aesthetic Conference

2012-01-05T21:14:00.000-08:00

I'm privileged to host, along with Adam Fitzgerald, a boutique conference next week (9-12 January) in New York City. The notion is that a small serious-minded group will take turns leading aesthetic experiences using the resources of the city and guiding discussions which relate those experiences to seminal critical texts. The presenters are:

Adam Fitzgerald, poetry, Columbia University and the New School, on de Kooning, Ashbery, Sontag, and Adorno
Michael Dennis, religious studies, Guilford College, on Rothko, dance, shared spaces, and the aesthetics of action
Sara Ceilidh Orr, Russian literature, The Ohio State University, on Bresson's Pickpocket
Jamie DeAngelo, art history and painting, The Art Institutes, on paintings in the Frick Collection
Stephanie Borger, English literature and Christian theology, Eastern University, on Mahler's Ninth Symphony
Laurel Hart, photography and art education, Concordia University (Montreal), on photography, social media, and public space through an Instagram tour of Chelsea and the Highline
Ryan Miller, philosophy, Boston College, on artist, audience, and critic in postmodern relationship through Punchdrunk Theatre's Sleep No More and MoMA PS1's interactive performance installations
Zachary Pace, poetry, Sarah Lawrence College, on poetry and place through the poems of Ashbery, Graham, O'Hara, and Roethke at the Hudson River's Christopher St pier

Sarah Vanacore (creative writing, Moore College of Art and Design), Louis Chartrand (philosophy, University of Quebec at Montreal) and Anne Carpenter (theology, Marquette University) were invited but sadly could not attend this year. The complete schedule follows.

Against Bernoulli

2011-11-13T18:55:00.000-08:00

It would seem like a teacher's first job would be to not tell children falsehoods. I don't mean that we shouldn't simplify (the world is complex; without simplification nobody could ever learn anything). The difference is that while simplifications (F=ma, the civil war was over slavery) don't need to be discarded, but merely augmented, when moving to the next level of understanding, falsehoods need to be painfully discarded (anyone who has ever chronically misremembered a name, or taught a college class, can attest to the painful part).

I think physics has the best track-record when it comes to doing this correctly. Almost all progress in the discipline can be understood as broadening semantics [e.g. m=m0/(sqrt(1-v^2/c^2))] or adding new equations in to reduce the error terms (friction). Physics teachers even tend to be pretty conscientious about adding words like "assume no friction" to problems, and explaining that relativity doesn't model small-scale behavior all that well. Chemistry, on the other hand, is full of nonsense like the "Pauli exclusion principle"--though there at least the actual answers are the subject of ongoing research. Many biology teachers are apt to say almost anything to children, as Cracked points out in its usual vulgar, humorous way. Cognitive science is an entire discipline founded on a falsehood, but at least there it's undergraduates being misled. If you thought history books were mostly full of facts, read Lies My Teacher Told Me, which systematically dissects every leading high school text. The Cambridge Grammar of the English Language debunks the falsehoods English teachers tend to make up, which are especially insidious as the students are often correct before formal instruction takes place. It's difficult to worry too much about the quality of instructional methods when much of the content is just made up--maybe states are right to emphasize basic literacy and mathematics exams?

But really that's all just a roundabout way of leading up to my main point, which is that there's one thing physics teachers consistently get badly wrong: the Bernoulli Principle is not the primary source of lift for aircraft or propulsion for sailboats. Even the otherwise excellent Wikipedia page gets it wrong (and no, I'm not wading into an edit war). Even comprehensive overviews by real physicists, who don't work in fluid dynamics, get it wrong. You should read that article anyway, though: it shows how basically every mechanical principle is involved in generating the basic behavior of the simplest Bermuda-rigged (triangular) sailboat to windward. I don't know why sailing examples aren't used more in physics and engineering classes; they're simple and engaging yet wonderfully broad, and would make great capstone projects.

But back to the subject, and to make things worse, most of those who realize Bernoulli's Principle can't be doing most of the work in low-speed, highly-loaded wings tend to assume, because of the observable downwash and angle of attack, that it's a simple Newton's Third Law situation: that's wrong too. If you've made it this far in my rant, go read that article: you won't regret it. It's not really any harder to explain than Bernoulli's Principle, and it's actually right. It might even be easier to explain, since smart kids won't stop asking what makes the air molecules get to the other end of the wing at the same time (spoiler: nothing, and they don't). Not only do physics teachers at all levels incorrectly use Bernoulli to explain everyday lift situations, they fail at explaining the (true) principle itself. The pressure change causes the velocity change, not the other way around (what magical force did you think caused the acceleration to the higher velocity?). At least Wikipedia gets that one right.

The correct theory of aerodynamic lift also explains why hydrodynamic lift increases when the boat is heeled and the sails are trimmed such that the boat experiences 3-4deg of weather helm, corrected with the rudder: it's like putting the flaps down a notch on an airplane. Of course, like on an airplane, this also increases drag, which is why there's a sweet spot when the boat is optimized and in the groove. In fact, some boats get all of their windward keel force from the rudder, with the keel just a skeg with a heavy torpedo for heeling moment. If keels worked by pushing water sideways, race boats would still have long keels, instead of the deep efficient wings you see. I wonder if a free rudder at the right distance aft of an efficient keel experiences slot effect?

Bernard Lonergan

2011-11-13T14:28:00.000-08:00

If you've wondered about the title of this blog, it's an allusion to the birthplace and philosophical method of Bernard Lonergan, S.J., a 20th century transcendental Thomist and polymath. I'm especially indebted to Lonergan both because his insights pulled me out of a sort of dry, analytic Scottish Realism (c.f. Alvin Plantinga) and because the scholarly community comprised by his friends, many associates, and students (and their many, many students, and their students) has been a tremendous intellectual support and inspiration. I've thus attempted to do a little scholarly work on Lonergan myself, should anyone be interested.

Papers which attempt to further elucidate Lonergan's work:

1. An attempt to diagram and make explicit the levels of consciousness Lonergan describes, presented at the 2011 Fallon Memorial Lonergan Symposium of the West Coast Methods Institute at Loyola Marymount University in Los Angeles: abstract, paper, slides, audio*. I would like to continue this line of research in a more rigorous fashion, perhaps with further conference papers or an article for Method: Journal of Lonergan Studies, but I'd like to complete a journal article on chemistry (below) first.

2. An attempt to understand Lonergan's theory of art in dialogue with modern and postmodern visual and performance art and the critic Susan Sontag, presented at the 2011 Lonergan on the Edge graduate student conference at Marquette University: abstract, paper, slides, audio*. This paper is in fairly rough shape; I think Lonergan's theory of art and art's importance for philosophy are both under-emphasized, but I don't have much/any scholarly background in aesthetics and it's not an immediate priority.

Papers which attempt to apply Lonergan's work to other fields:

1. An attempt to explain why chemistry and physics are separated by more than convention, without appeal to downward causation, presented at the 2011 Philopolis applied philosophy conference at McGill University in Montreal: abstract, slides, video*. This is my most active area of research right now, as I work my way through enough physical chemistry and philosophy of science to submit a journal article to HYLE or Foundations of Chemistry supporting my thesis.

2. An attempt to explain why biology and chemistry are separated by more than convention, without appeal to downward causation, presented at the 2010 Philopolis applied philosophy conference at McGill University in Montreal: abstract, slides, video*. If the similarity in titles didn't give it away, this is the paper that convinced me that philosophy of chemistry is where the action is.

3. An attempt to use Lonergan's methods to dispel the force of Kierkegaard's argument in Philosophical Fragments against the Augustinian synthesis of faith and reason, presented at the 2002 St. Peter's College undergraduate colloquium in philosophy: paper, slides*.

And then some undergraduate work applying Lonergan to Hegel and Einstein.

* I hope to add links to the live presentations when I get them from those who did the recording.

Puppet & Cross-Cutting Concerns

2011-11-13T11:16:00.000-08:00

One of the hardest things about managing a Puppet installation for a complex infrastructure is handling cross-cutting concerns (servers that need to get a piece of configuration data based on the cartesian join of location, role, project/cluster, etc.). Currently supported ways of doing this all have serious drawbacks. You can do all the assignment manually in LDAP or some other directory, but that defeats the point of automation. You can use extlookup, but it only supports a single hierarchy of overrides. You can just have separate puppet instances for each cluster and reuse your modules, but what if you have servers that are in multiple clusters? (Such things are often dismissed by small shops with a single project or large shops with many servers in each role, but many real medium-size businesses with multiple products have such problems, and more or less inevitably so.) I think Puppet needs to seriously look at decorators or some other method of handling this organically, which would require some aspect-oriented-programming support in the language.

Black Boxes and Overfitting: The Twin Cases of EBS and CDOs

2011-11-13T07:59:00.000-08:00

I think there are some instructive comparisons to be made between the motivations, technologies, and failure modes of Amazon Web Services Elastic Block Store (EBS) and investment banks' collateralized debt obligations (CDOs). With luck, elucidating those similarities will help lessons learned in each area mitigate risk in the other and maybe even help technical workers better understand financial risk and financial workers better understand technical risk (in the spirit of technical debt).

1. EBS and CDOs were both born from the insight that sharing can reduce risk. Before EC2/EBS, companies could either bet on high usage, with the attendant risk of having lots of capital tied up in non-productive assets if traffic was lower than expected or efficiency was higher, or bet on low usage and fail to serve customers if traffic was higher than expected or efficiency lower. EC2/EBS uses a shared infrastructure, so capacity projection can happen at the (hopefully more stable) level of the internet as a whole, with resources dynamically allocated to whoever needs them at a given time.

Before CDOs, financial institutions tended to have large exposures to the risk of those regions or industries where their market presence was highest, and since particular regions and industries tend to have more volatile economic profiles than the world as a whole, they either needed to carry excess non-productive (reserve) capital to offset that risk, or potentially go bankrupt in a sectoral downturn (e.g. Dustbowl banks during the Great Depression). CDOs allow each financial institution to package risk in a standard way, so that they can buy and sell whatever portions are necessary to achieve an optimum risk profile within their budgets, rather than being hostile to the vagaries of the markets in which they operate.

2. EBS and CDOs are both marketed by trusted vendors of other products who ask for more faith. Whatever you personally think of them, both Amazon and the major investment banks (Goldman, Merrill, Morgan, Bear, Lehman) were broadly trusted for all kinds of transactions before they started to push EBS and CDOs. Not only unrelated kinds of business (retail brokerage, Christmas presents) but also more direct precursors in the sharing of risk: asset backed securities and S3.

3. EBS and CDOs are both driven by concerns about time to market, transaction costs, and labor costs. In addition to chasing lower risk, businesses also want to cut expenses, and having standardized structures in which somebody else does the legwork that you can buy and sell on the open market at any time seemed like a great way of doing that. Tech companies could stop hiring sysadmins and purchasing agents. Banks could stop trying to grow and acquire their way into more diversified markets. Risk and computer time could be bought and sold whenever the business required, rather than waiting for some difficult logistics or paperwork chain to swing into action.

4. EBS and CDOs are both an attempt to make features from an old paradigm available in a new one. I can't really think of anyone who would deny my first three claims, but this one is more subtle and possibly contentious. Cloud services promised a new world of abstraction, ephemerality, and explicit guarantees. The upside made companies desperate to move their web operations onto the platform, but their entrenched data models didn't fit into Dynamo's simple key/value paradigm and their legacy databases didn't replicate well enough for ephemeral storage to be sufficient. Amazon wanted to please their customers (and increase profits) so they put a lot of duct tape and bailing wire around iSCSI, DRBD, and LVM and called it EBS. All of the cloudy resource sharing, but now with permanent* block storage to accommodate legacy databases, carved out of some set of disks transparently mirrored and replicated behind the scenes. Users then began to rely on that leaky abstraction.

Investment banks did something similar with CDOs. In the new world of high frequency trading and complex computer models, once the mortgages were standardized into pools buyers could theoretically have bought any cross-section of a deal that they wanted, tailored to their needs, and subsequently saleable on the open market at whatever appreciation or depreciation then applied, just like equities. Less sophisticated investors, however, like retirement funds, weren't equipped (or potentially allowed by statute) to run complex computer models on each segment of a transaction. Instead, they demanded large blocks rated by the agencies (e.g. Standard & Poor, Fitch, etc) at discrete ratings (e.g. AAA or investment-grade). So, like Amazon, the investment banks listened to their customers and their bottom lines and delivered. Standardized mortgage pools were tranched into more complex structures that allowed large investment blocks to be declared AAA.

5. EBS and CDOs both provide incredible value in good times--which are most of the time. Baron Schwartz explains why virtualization has a very steep normal/worst-case performance curve, and why it's difficult to even find out what the worst case performance is. Given that, the median performance will be well above the average performance, which makes cloud service seem like a very good value. That's even more true of EBS, where the additive variation occurs at each level of the stack that is virtualized (cpu, network, storage), and thus has even fatter tails.

CDOs worked similarly: having achieved a AAA rating with better yields than treasury bonds or similarly-rated corporate debt, during non-severe-recession years there was no way to detect hidden risk, and only the higher income stream was evident. Public pension funds and private hedge funds both looked flush with cash, but were really just lucky.

6. EBS and CDOs both created environments where users were hyper-aware of luck at a micro-scale, while completely ignoring it at a macro-scale, so hedging strategies were actually counter-productive. EBS users are so aware of randomness in the loading of particular parts of the infrastructure that they often create new volumes, test them, and discard them if performance is poor. They devise best practices around hardware failures. As that first article explains, however, performance failures turned out to be correlated, and as both AWS itself and its users tried to find enough working sections of the system to remirror their data, more and more sections became overloaded and failed. In fact, the more users tried to spread data across volumes, the more pain they felt. User behavior in attempting to account for local risk actually changed the way the service was being used enough to generate increased global risk.

CDOs work similarly. If particular mortgages and bonds weren't risky, there would be no point in securitizing them. It turns out that securitization is highly dependent on the correlation factor of the underlying assets, however, which of course turned out to be higher than expected. So the creation of CDOs lowered the expectation of risk, but didn't actually lower risk, meaning that lots of perceived value was destroyed when they crumbled. Furthermore, as banks tried to further hedge their risks with credit default swaps (CDS) on those CDOs, sellers of credit default swaps, like AIG, would go bankrupt in a bust even if they had no exposure at all to the original loans. Worse, the very creation of CDOs allowed credit to flow more effectively, creating a bubble in housing prices which changed the assumptions on which the CDOs were based.

7. EBS and CDOs both create abstractions which are not amenable to performance observation or prediction in crises. Much of the pain suffered by customers around EBS is because the abstraction's developer contract is so broad and opaque. Amazon can restore an S3 bucket sans one file, and you have most of your life back. Amazon can't restore an EBS volume without a certain stripe because your filesystem won't mount. Conversely, the end user doesn't know when snapshotting and moving to a new volume will help (because the contention is purely local and random) or hurt (because the whole system is experiencing pain). Since neither party knows what data is actually where, neither party knows what's really going on.

CDOs suffer a similar problem with lack of transparency. Since the companies who deal with mortgage customers don't actually own the loans, and in fact many entities may own shares of a single loan under different contact terms, preventing effective loan modifications. CDO default correlations are the outputs of models with enormous numbers of variables, each with uknown distributions, making it impossible for different parties to agree on valuations during periods of market volatility--which is why CDOs were so illiquid during the crash, and basically solvent banks had to accept government assistance to pay their daily bills.

Lessons learned? In both cases, some major firms, like Netflix and Goldman Sachs understood the underlying architectures well enough to avoid using those parts which were not sufficiently transparent for their needs, unlike Reddit and AIG which suffered mightily from failing to do so. The obvious lesson is that if you're making money but you don't really understand the underlying models that power your business, you're probably unwittingly taking on a lot of risk, perhaps at someone else's gain. TANSTAAFL, as I believe Smarter Travel Media may have found out in the online advertising department.

The more effective your business is at optimizing to a set of constraints, the more important it is that you understand the forward validity of those constraints.

On the provider end, realize that most of your users don't know what abstractions are actually helpful. You have to trust your gut to come up with good ones--and you'll remember that AWS didn't ship with EBS, and no other cloud provider has anything like it. I wonder why? Similarly, Goldman hedged their CDO risk instead of banking on it, and most investment bankers are incentivized to ignore their instincts on risk. If you're pressured by customers/profit into offering an abstraction that your gut knows is wrong, prepare technically and legally for assured future pain.

Why You Shouldn't Use MongoDB for OLTP

2011-11-11T20:57:00.000-08:00

Why is MongoDB a popular option?
1. The developer libraries are really polished and standalone systems are easy to install. Many database projects are developed by extremely talented people who think that because writing a connection library is easy compared to the database internals, end users would really just want to write their own custom stuff anyway. Mongo and MySQL succeed in no small part because they fully recognize that developers are expected to have a working prototype yesterday.

2. It seems to offer the scalability and power of HBase/Cassandra with the ease of use of CouchDB/MySQL. Of course you can't get something for nothing, but "something for me now" at the cost of "maybe something for somebody else later" is often not an attractive tradeoff. Additionally, 10gen try really hard to gloss over/ignore/propagandize over those tradeoffs.

3. Mongo is the best way of solving the unsolvable problem of running a high volume Web 2.0 OLTP database on EC2, due to the horrific storage latency of EBS.

So let me take those in order.

1. That's just true, and is a lesson that lots of infrastructure developers should bear in mind. It's what made Steve Jobs rich. Usability matters, whether we like it or not. One thing organizations can do to keep this from skewing their judgment too much is to take on the DevOps model, where the production sysadmins are right in there with the prototyping developers, and can hopefully help them get environments up and running quickly while also accelerating the production sysadmin's understanding of the platform. Win/win. Of course in purely exploratory use cases this may be the only thing that matters, but I think part of the lesson of the agile movement is that there basically aren't any of those. If they wanted the prototype yesterday, they'll want the production implementation yesterday too, so you'll reuse the prototype.

2. This is the big one. Alex Popescu passes on an anonymous rant that summarizes the issues. The rant itself may be a hoax, but the design complaints echoed by Alex's commenters are no less trenchant for that:
a. Like MySQL, Mongo's performance benchmarks are all done with the safety features turned off (immediate writes, writeahead log). You either get (drastically) less performance than you expected when you flip the production switch, or you cross your fingers and pray.
b. Like MySQL, Mongo is optimized for reads (everything in RAM) and not writes (global write lock!)--except writes are the hard problem, writes are the problem that doesn't occur in dev and is hard to simulate in test, and reads can usually be papered over with caches.
c. Like MySQL, Mongo relies on a master/slave arrangement for availability. Anyone who has dealt with MySQL replication at scale, like say Mark Callaghan, knows this is a nightmare. Percona estimates the approach at only three nines of uptime. Making replication truly crash-safe is a really hard problem. Making sure that you have a known window of data loss during slave failover is also a hard problem. Both need to be baked in from the beginning--databases need to be fundamentally clustered (this is of course why things like HBase/Cassandra/Oracle RAC are harder to get going in the first place).
d. Like MySQL, resharding is a bolt-on feature of Mongo, instead of being baked in from the beginning. Given that this is basically the hardest thing to get right in a clustered database, that's a recipe for disaster.

In general, 10gen like MySQL looks good on Wiki pages and RFP responses because all the boxes are checked, but fails in real life where you want all of the boxes checked at the same time. Not only is that borderline dishonest, it's a terrible way to write software since it's almost impossible to fix later on. Beyond that, the CAP theorem is a harsh mistress, and as the people who actually write this stuff keep telling you, the only way to make an accurate tradeoff is to know your data really well (what's the locality for reads and writes, does the distribution work well for bloom filtering, etc.). When Mongo offers a schema-free solution that offers great query performance without knowing anything about your data, you can be sure that CAP is going to bite you later in ways you didn't expect.

3. This is also probably true. EBS is a lie, which is to say inevitably brittle, and like with Mongo and MySQL, it's a design level problem that implementation fixes won't help. So your options right now are:
a. Magically force your data into a pure key/value system and use Dynamo.
b. Use Mongo (or RDS if your volume is low) and live with lost data and downtime. With latency like this you can't afford to doublewrite or synchronously replicate anyway. If your business is FourSquare, that might be ok.
c. Buy servers.

A National Disgrace

2011-11-11T18:59:00.000-08:00

So my last blogging venture gradually became more and more focused on my outrage over government torture (and murder) abroad and wiretapping at home, until Glenn Greenwald started writing about both at a level of quality I couldn't touch, so I left him the field. Those who think that political blogging is light on research or unnecessarily shrill should spend some time with his writing. He has built, through original research and public documents, an incredible case on the systematic bipartisan abuse of government power to take life and liberty. Anger can be righteous and when the outrage continues without apology (and indeed with ongoing increase and coverup), that righteous anger should grow rather than ceasing.

So I've no intention of returning to blogging about that full time, but Andrew Sullivan reminds me that people are still being murdered in our name. If you haven't joined the National Religious Campaign Against Torture, please do so. If you don't think we're torturing people, read Glenn's books. If you think that what we're doing isn't torture, exactly, ask yourself why we executed enemy soldiers for those same actions in WWII. If you think torture is ok, may God have mercy on your soul.

The Value of a College Education

2011-11-11T12:51:00.000-08:00

By request of ZT205, and because Tyler and Alex can't seem to shut up about it either. Really, this should be entitled "a list of things you should do/learn by age 21" because the line between college and high school is blurry (APs, dual enrollment, community college) and I'm not convinced that the present U.S. (Germanic?) system of post-secondary education is really the best fit for teaching these things anyway. Nonetheless, most American kids will be best placed to learn these things while in college, so the headline is appropriate.

1. Quantitative reasoning. To make decisions in the modern world, you need logic (rigorous, i.e.. mathematical/symbolic/formal), computability/optimization (linear algebra/discrete math or similar), statistics (ideally including non-parametric stuff, which invariably has calculus and statistics 101 as prerequisite), and rates of change (calculus 1). With the exception of statistics (where the typical 101 course is a disgrace and should be taught in a semester of high school) and potentially logic (where some schools make informal logic a prerequisite) these are all 101-level classes. Typically, however, even A and B students really need to take the follow-on class before they're truly solid on the grounds of the 101 class and able to apply it. The ability to understand and apply this material qualifies you to understand and build models of the world--without which you cannot make rational decisions in business, economics, politics, or science.

2. Qualitative reasoning (hermeneutics). In order to function in a pluralist, inter-disciplinary world, you need to have a rich theory of dialogue. You need to understand how positions and language are shaped by tradition, and how those traditions can talk, even argue, with one another. This implies moving beyond reflexive traditionalism and faddish "everything is shiny" liberalism. It implies being able to read and write well, without which careful dialogue is impossible. It necessitates some understanding of change and context in history. Reading and writing in a second language can certainly help with this, but I'm not convinced that it's absolutely necessary.

I am deeply concerned that the number of students who are getting both (1) and (2) in a serious way is far too small, and that's a main cause of why people are able to be hoodwinked by fallacy and ideology, and economics and theory types are unable to speak meaningfully to one another. In addition to the democratic value, the combination of these skills also has incredible economic value in that together they imply the ability to build a model and negotiate its application to a foreign world--the heart of decision making in any business. Thus I heartily endorse schools like MIT which insist all students have a foundation in both kinds of reasoning, and I'm skeptical of those like Brown without non-major requirements. Brown gets away with it because most of their students are at least interested in qualitative reasoning and had calculus in high school, but it's no model for other schools to follow.

3. Empirical method. Learn at least one discipline's method of collecting data and modeling explanatory relationships. Understand how the basic method can be applied to subject matter outside the traditional ones of the discipline, as well as its limitations in doing so. This means taking courses through the 400-level in some kind of sequence, and rules out things like "pre-med" which only feature one course beyond the 101 level (organic chemistry). This will equip you with an important method to approach unknown problems, even if not always the optimal one, and should be to a high enough standard that you can apply the method professionally (e.g. engineering) or undertake original work in it (e.g. graduate school).

4. Professionalism. Learn to build relationships and get things done. Graduating college without a single professor you know personally and can rely upon for a reference is disastrous. More particularly, find out how the standards for relationships and getting things done in a particular field actually work. That means you need to attend academic conferences and/or engage in internships. Far too many students graduate without any realistic notion of how people who employ the method they've studied actually impact the world, and get into positions to do so.

I'm firmly convinced that students who do these four things can succeed, no matter what they major in. As the commenters on Tyler and Alex's posts point out, attempting to figure out what majors will be in demand when you graduate, or going back to school for several years each time the needs of the economy change, are futile endeavors. I stand behind the adage that education is about learning to think. Where students get into trouble is by failing to learn to think with sufficient breadth, rigor, and practicality. Without breadth, i.e. the ability to transfer learning, you're too easily stuck in a niche that may evaporate or never exist in the first place. Without rigor, you've just learned a bunch of fancy words that don't really give you any capabilities you didn't have before. Without practicality, you'll never meet the right people or actually accomplish the work to enter your chosen profession or make a difference within it. I wish academic advising systems could somehow be made to take this seriously without being hijacked by academic bureaucracy or narrowness.

None of this, of course, suggests that some professions won't have higher payscales than others, but I think that's ok. If you want to do something with immense non-monetary rewards (puppetry, academia) you'll have to be among the best if you want monetary reward also--no sense whining about that any more than the investment bankers get to whine about their boring jobs and lack of free time. Or that most Americans wouldn't be happier with less stuff and more free time. But I'll save consumerism for another post.

5. Make friends who will last a lifetime. At no other time in your life will you be surrounded by people with such similar interests and so much time to devote to developing friendships. You may marry one of them, or marry one of their friends who you meet later on. Many people have trouble developing friendships in high school due to the maturity level of all parties. After college, there is a high chance that you will be in a nearly-unisex environment (military, pink collar, engineering), or one where everyone goes home to their families at night without much socializing. Don't waste this time on drinking (I've nothing against alcohol, but the friendships forged on it won't last well after it) or vapid friends who you enjoy being around but don't challenge you to grow.

Extemp Sources

2011-11-09T16:36:00.000-08:00

The Economist, The New York Times. These are the newspapers of record. If they print it, it is by definition news and fair game for an extemp question, and you never need to justify or contextualize citations to them. Which doesn't mean they're always right. There's no excuse not to have every Economist article from the last twelve months in the tubs. If you get behind on the Times (which does require a bit of judgment, or you'll end up with thousands of daily updates which don't add much information to ongoing stories, and lots of one-hit-wonder stories which would never justify a question), at least cut the news analysis and Sunday Review articles which synthesize the news at a higher level. Some would put the Washington Post in this category, but most of its coverage duplicates the Times, and much of what doesn't is overly-specific inside-the-beltway stuff. It is nice to have another broad source of record, however, so I advise cutting the politics, economy, and policy sections as a compromise.

The IHT used to be a fine source before it became merely the global edition of the Times. The Christian Science Monitor is also a very good paper, but can sound odd to judges and only rarely has reportage that the Times and Post don't. Other domestic papers have really cut back their national and international bureaus and tend to almost exclusively run Times, Post, and newswire content. Time Magazine, on the other hand, is really trying to become the domestic Economist in the wake of USN&WR's demise, and at least 2/3 of every issue is worth cutting. The quality of reportage has gone up tremendously, and it's a great domestic analysis source, even if judges don't all realize it yet. USA Today's ratio isn't that good yet, but it does have really great articles on occasion, so it's worth a scan. The Oval and judicial are its best sections, though blog and print content are not well disambiguated--but that's increasingly true of the Post as well. I don't think blog content is categorically uncitable, given how impressive its authors credentials can be, the depth of analysis going on (at least occasionally), the way it's interpenetrated print media, and how little editing and fact checking goes on in print anymore--but citing them as if they're part of the print publication is iffy, judges can be turned off by the idea of citing a blog, and it can be difficult to teach students to differentiate useful and worthless blog posts. Presumably this problem will get worse before it gets better. National Journal is also an excellent and underappreciated, though voluminous, source. Der Spiegel is probably the best international paper to cut...most of the British ones are tabloid-esque (though they occasionally run really strong stories that are undercovered in U.S. papers, e.g. Robert Fiske), the Israeli papers can be hard to make sense of from an American political perspective, and the Australian papers tend to be either provincial or repetitive. The South China Morning Post used to be amazing before China started stepping all over press freedom in Hong Kong.

Beyond the general sources of record, it's also worth collecting background information, partisan perspective, in-depth trend analysis, primary source data, and additional material on trending topics. The NYT Backgrounders, Economist special reports and predictions, Congressional Research Service, and Cato Institute Policy Guide are all fantastic repositories of synthesized background material on topics in the news. A speech that can discuss how events do and do not conform to received expectation (and why) is a strong speech. Cato, of course, is a libertarian thinktank, which leads to our next topic: partisan sources. Partisan sources, as long as they retain a modicum of honesty, add great color to a speech. They usually cover infighting more than mainstream sources, and they can also be used to explicitly cite what liberals or conservatives might think about an issue. Mother Jones, The Week, The National Review, The Weekly Standard, Salon, and The New Republic are all exemplars of this genre. Some of them are more honest than others. Cutting each one every week would be overkill--sucking up not only cutting time, but precious time scanning the articles before speaking--but grabbing a selection of articles on rotation adds depth to your files. Bringing a more (or less) academic perspective, with varying degrees of partisanship, but generally longer-term trend analysis, are the monthlies, quarterlies, and book reviews. Foreign Affairs, Foreign Policy, The Atlantic, The New Yorker, The NY Review of Books, The London Review of Books, and Vanity Fair are all great sources in this category. Many of these will only have 2-3 political articles in each issue, so make sure students aren't cutting music reviews. Articles from these publications also tend to be a lot longer without the explicit internal structure of the backgrounders, so highlighting is key and discussion at practice can also be useful. This genre also includes more specialized sources like Wired, Columbia Journalism Review, New England Journal of Medicine, Harvard International Review, Washington Report on Middle East Affairs, and the lamentably deceased Far Eastern Economic Review.

Primary sources also make great extemp evidence, but it can be difficult to sort through the volume available. This is where websites like StratFor, the Center on Budget and Policy Priorities, Calculated Risk, and Marginal Revolution can be invaluable. They reliably pick out, link, and contextualize government reports which underlie the news--but these sources are voluminous, and not especially citable in their own right. I think the best plan is to print both the abstract of the government report (not the whole thing) and the blog post, and file them together. Marginal Revolution also includes a lot of great material for intros (and lots of technical economics and totally irrelevant things), and also does a good job of linking to the best articles from the Financial Times of London, thereby maximizing your ten free articles per month.

It's also worth bookmarking collection pages from reputable sources on trending topics, even if it's not worth cutting that source more generally, e.g. the BBC on East Africa or the Arab Spring. This is even worth doing with publications that you intend to cut fairly completely, in order to make sure you don't miss something (e.g. the Economist on the EU sovereign debt crisis or the Post on housing and the recession. If students are having trouble understanding a topic (and they won't make time to read a book and practice time is limited--you know, the usual) Reddit has an Explain Like I'm Five section which can be of help (at least as a starting place for discussion--as with any forum, it has its share of errors, though crowd moderation is surprisingly effective).

The Tension of Crisis

2011-11-08T11:48:00.000-08:00

We all face setbacks--but when is a setback a crisis? Well, look at corporate bankruptcy: when a company faces a severe enough setback that it cannot pay its creditors, it is insolvent. If creditors do not believe that the business can regain profitability in the future (after the sunk cost of the principal writedown) then the business will have to undergo Chapter 7 where its piece-parts are sold rather than a Chapter 11 restructuring of debt and operations. That guarantees a large loss for all parties because of the first principle:

1. Everything is worth more as a going concern. The difference between net balance sheet assets and valuation (what someone is willing to pay for the company, either on the margin as stock equity or wholesale by acquisition) is goodwill. Excepting some weird special-purpose-vehicles (i.e. tax avoidance, regulatory avoidance, and fraud) and highly unstable situations, goodwill is always positive (if it were negative, a rational owner would immediately sell all assets on the open market and fold the company). Even if it were par, that would surely be a lower risk, and thus preferred, strategy.

Since investors in crises are already facing losses, they have strong incentives to retain their investments as going concerns lest those losses rapidly multiply. Thus arises a corollary:

1.1 Investors should attempt to manage the scope and timing of losses so as to maintain the perception of future profitability and ability to pay, lest they lose tremendous value in goodwill.

Markets that clear, however, are better than (non-)markets that don't. Assuming both parties eventually want or need the market to clear (thus excepting e.g. heirlooms), they're eventually going to take the spread haircut anyway, in a zero-sum fashion. Trade itself, however, is not zero-sum, and higher velocity means more growth. Markets that don't clear are thus delaying haircuts and causing stagnation. So the next principle:

2. Functioning markets close the spread and clear on the margin.

And in a crisis, of course, with the prospect of lost wealth and deleveraging already hurting the economy, growth is desperately needed as the cure. So this principle has its own crisis corollary:

2.1. Writedowns in a crisis should happen as soon as possible.

You can probably already see the tension between 1.1 and 2.1, but let me spell it out further. Most obviously, human beings generally react to rapid change, especially negative change, with denial. First, however, there are a number of potential private advantages to being a late-mover in taking your haircut. Second, purchasing and/or growth may resume in the meantime lessening your haircut due to your firm's greater worth as a going concern. If your investment truly is the worst, you may be able to delay with smoke and mirrors until somebody else goes broke, and increased trading then improves consumption and credit. If your investment isn't the worst, how can you prove it when the others in your class are blowing smoke?

So if human nature tends to favor 1.1 over 2.1, to collective detriment, what can we do about it? Good Chapter 11 procedures help, as they make 2.1 look less scary (currently good for U.S. corporations, not so good for banks and other financial corporations, absolutely terrible for E.U. governments). Good beginning accounting and publication rules can help with transparency, though introducing new ones during the crisis (e.g. bank stress tests) doesn't seem that helpful as the market doesn't have existing expectation mechanisms around them. Bailouts and guarantees seem flatly counterproductive.

I would suggest that, as with unions, step one is to see if the market can be self-correcting. Let the speculators do their thing--if you don't want to be speculated on, take on less leverage in the first place. I think most of the objection to speculation is that it's unseemly to make money off of others' pain (though you can donate it to charity to take the edge off) but they're actually doing the incredible public service of correcting group bias.

So like Iceland, I think we should just let the banks fail. Goading them into mergers just moved the instability up the chain, and government guarantees and bailouts just move it up further still. Eventually the piper will be paid, so we may as well get on with it.

Organized Labor: Where from Here?

2011-11-05T20:36:00.000-07:00

My mixed feelings on unions are presumably a result of my lack of a coherent theoretical framework for political association (I'm a Catholic with libertarian leanings) and unwillingness to overlook the concrete facts of history. So first, the reasons that unlike many libertarians (and historically, many Catholics) I'm unwilling to dismiss labor unions as an important part of the body politic:

1. Firms exist because it's inefficient for every individual producer to manage his own contracting practice. But isn't that just what individuals have to do when they decide on terms for selling their own labor to firms? So having an agent who specializes in that process just makes sense. This is why engineers and scientists achieve such poor financial returns on their crucial labor.

2. Individuals aren't well-placed to negotiate for changes with large externalities for other workers, like safer machinery.

3. Lots of services, like job training, testing, insurance, etc, may be better handled by industry-wide unions than firms who may employ lots of people, but only a few in that specialty or for short duration. The entertainment and construction trades are good examples of this.

4. Individuals who are in difficult personal situations are both the most dependent on continued employment with the same employer (search costs and medical insurance both being big deals for them) and the least able to negotiate from a position of strength for such employment. Unions provide representation and pre-agreed rules to govern such situations.

But current forms of labor organization carry significant costs as well:

1. They attempt to cartelize labor, benefiting their membership, and sometimes achieving a better share for labor overall, but at the expense of non-members, including immigrants (who tend to be the poorest among us) and workers in other countries much poorer than our own, who suffer from truly grinding poverty.

2. In attempting to standardize conditions and prevent management abuse, they often prevent management from doing the right thing in particular cases, like being more lenient or rewarding workers for performance. They also tend to draw sharp lines between labor and management that can prevent natural gradual increases in responsibility.

3. As with any organization, unions are subject to their own public choice analysis. Minimizing costs for hiring new workers maximizes their potential membership, while back-loading compensation gives members strong incentives not to leave the firm or industry and thus the union. Often, however, this is not in the best interest of the worker or the employer--as evidenced by difficulties compensating new workers sufficiently to induce entry and early retirement packages designed to shed costs while pushing out the best workers who are doing what they love.

4. Formal bargaining tends to set long-term inflexible strictures on the basis of short-term observations, leading to situations like Big-3 and state government pensions whereby unsustainable promises are made to ease present frictions. Both the union and management are incentivized to view the future with rose-colored glasses, at the expense of retirees who end up making decisions based on false information.

So what's to be done? Maybe more unions could be like those in entertainment, who don't demand uniform compensation or delayed benefits, but do negotiate hard for total profit shares and provide other services to workers? Maybe arbitration clauses, already common, could be relied on more instead of over-detailed contract language? Most importantly, I'd like to see what arrangements companies and unions would come to without government guns trained on strikers or NLRB support for closed shops. I think this is one of those cases where we've yet to give the market a try.

The Intermittent STEM Education Panic

2011-11-05T19:58:00.000-07:00

There's been a lot of handwringing lately over the issue of whether enough U.S. undergraduates are majoring in science, technology, engineering, and mathematics/medicine (STEM) to support global competitiveness/continued innovation/economic growth/upward mobility/middle class life/tax base/etc. I don't have an easy answer, or even a hard one, but I do have a couple of completely unscientific comments and questions.

1. Waits for doctors are long; costs for doctors are high; many people want to become doctors. So why don't we train more doctors? Every year up-and-coming universities start new business and law schools, but there aren't any more medical schools and places than there were two generations ago. So getting in is a herculean effort, and the whole undergraduate process gets skewed. What's the constraint here?

2. In high school or even earlier, we need to do a much better job of teaching math-types to be people-literate and people-types to be math literate. Engineers who aren't people-literate can't negotiate for higher wages, or sell their new ideas themselves, or write grant proposals to fund their new ideas, all of which would create innovation and bring more candidates into engineering. How can businesses claim with a straight face that there aren't enough STEM graduates when they receive such a small share of revenues in companies highly dependent on their output? Similarly, however, people people aren't well placed to understand the value of engineers and innovations in the general or the particular, or do small bits of rigorous analysis for themselves, if they are innumerate. That most undergraduates are terrified of logic/math/statistics/computation and have no idea how to use them is nothing short of a travesty. Many ho-hum humanities/business/communications/journalism majors are pushed there from fear, not pulled there by a sense of vocation.

3. If you read the hand-wringing, you'd know there's a lot of complaint about busy-work, especially labs, that eat time and cause pain without really teaching anything. If you read the article about medical school numbers, you'd know that while overall numbers are flat, female numbers are up (and thus male numbers are necessarily down), and that holds in undergraduate education as well. I suggest the connection is that, for whatever reasons that unlike Larry Summers I will not speculate on, women in general seem to have a higher tolerance for homework stress just as men in general have a higher tolerance for test stress. So as we shift more grading onto homework, we improve women's educational outcomes and worsen men's. It turns out that the real world involves elements of grinding away and elements of high stakes stress, but mostly revolves around more dynamic situations where intelligence and effort both play a role (like a science fair project). Students who understand the material (as proven by test scores) should not be required to do homework. But students who prove their mastery of material via homework and projects should not be subjected to high stakes tests. Why is it so hard to accept that we're just looking for proof of mastery, however that can be demonstrated?

3. The education system systematically fails to deal with differences in rates of learning. Students who learn rapidly are pushed into advanced classes, where they initially do poorly because they lack the background and then are again bored when they catch up. Students who learn more slowly are forced to repeat classes or entire grades where they again find themselves lost and behind. If tracking means forcing students into patterns of classes organized by vocation or ability where later change is difficult or impossible, then it is an injustice. But if tracking just means putting students in classes where the rapidity of instruction matches the speed of learning, then not doing so is an injustice. In STEM fields, topics build upon one another, and a student who has not had the proper opportunity to learn the basic material has been done a great disservice.

4. This deserves its own post, but the messed up compensation system for teachers must take some of the blame. Instructors, both at the HS and college levels, are paid nearly nothing when they start out with college loans, mortgages, small children, and lots of compounding yet to go on their retirements, and are paid much more at the end of their careers (setting artificially high pensions) when they are still in basically the same positions reaching the same number of students. This is especially painful in STEM fields where graduates can easily find work in industry, because the best teachers have a strong financial incentive to leave.

5. An artificially high percentage of intelligent, numerate people are being pulled into software and finance. Partly this is because profit margins are high enough in those fields to support higher salaries, but it's also because talented people can gain employment in those industries without acquiring painful credentials. Any industry populated by lots of dropouts (i.e., who bear the selection effect of further schooling without its costs) is probably doing something smart. If you need an M.S. to employable in STEM, why not combine your B.S. with a J.D. or an M.B.A. instead and make a lot more money? Why on earth do H.S. teachers need masters degrees to advance in salary when the educational theory they're taught makes no demonstrable difference to the students and any additional subject knowledge will never be taught in H.S. or could be gained in much lower cost ways? I don't doubt that the Ph.D. is a useful process for most researchers, and that the bachelors is typically a useful precursor to the Ph.D., but any time you see basically no allowance for alternative paths you're closer to a cartel than a productive industry.

Welcome to my resurrected blog

2011-10-15T14:30:00.000-07:00

Mostly philosophy, but probably plenty of debate, sailing, computer science, folk dance, and current events.