Fire-and-Brimstone Places

Monday, March 11, 2013

Sam Harris, in an attempt to "give you some sense of what we are up against whenever we confront religion", writes of what he calls "the fireplace delusion":

I have discovered that when I make this case [i.e., that burning wood is about as dangerous as smoking --ed], even to highly intelligent and health-conscious men and women, a psychological truth quickly becomes as visible as a pair of clenched fists: They do not want to believe any of it. Most people I meet want to live in a world in which wood smoke is harmless. Indeed, they seem committed to living in such a world, regardless of the facts. To try to convince them that burning wood is harmful--and has always been so--is somehow offensive. The ritual of burning wood is simply too comforting and too familiar to be reconsidered, its consolation so ancient and ubiquitous that it has to be benign. The alternative--burning gas over fake logs--seems a sacrilege.
I personally have no trouble with the idea that a wood fire belches out toxic smoke, but then I don't subscribe to the popular dogma, common in some circles, of "natural good, man-made bad" (as if man is somehow not a part of nature).

I'd like to elaborate on an issue Harris has raised. When someone accepts something as fact on an arbitrary basis, he is pretending that certainty can be achieved without good evidence or sound argument. Hearing such a "truth" put to question will immediately put such a person on the defensive in the same way that hearing heresy will upset a religious person. Why? Because a bluff to oneself is being called and, on some deep, disturbing level, that person knows it.

So far, so good, but I take issue with Harris's closing sentence (See also postscript.):
Of course, if you are anything like my friends, you will refuse to believe this. And that should give you some sense of what we are up against whenever we confront religion.
If Harris is including the kind of people who bristle at any suggestion that fire smoke might be harmful as part of the "we" who are "up against" religion, he is wrong to do so. Traditional religion is just one manifestation of irrationality. Harris's fireplace cult is another. There is no room, in a fight to uphold reason, to allow anyone who permits himself to indulge in the arbitrary to masquerade as an ally.

Harris's attempt to shame irrational people into shaping up will fail in most cases. While people can change, the process of evasion that predisposes some people to accept the arbitrary in the first place usually becomes so entrenched that genuine introspection becomes nearly impossible. The fireplace delusion strikes me as less a teachable moment and more a diagnostic tool -- a warning that the person one is dealing with is not necessarily rational. (This is not to say that initial resistance to a new idea always portends the worst: There can be other reasons a person finds an opposing view disturbing the very first time he hears it.)

Harris does have a point, though I strongly suspect a bigger one than he realizes: Religion isn't the only thing an advocate of reason is up against.

-- CAV

P.S.: I also take issue with his assertion elsewhere in the piece that the recreational use of fire should be banned, but that is beyond the scope of this post.


Steve D said...

Well, though I agree with his (and your) main point about the struggle against irrationality, I for one take issue with his ‘scientific’ conclusion about burning wood. I’m one of those who bristle. Here's why.
This same issue comes up time and again during discussions about the health effects of second hand smoke (wood burning is analogous to second hand smoke if you think about it). It’s irritating since even people who should know better don't seem to be able or willing to pursue the problem properly and therefore end up treating cigarettes as these horribly dangerous objects which will kill you if you even so much as think about them. (kind of like how some people think about guns)
We need to take this one step at a time. The one actually doing the smoking is directing almost the entire output of the cigarette directly into his lungs; the one sitting across from him at the table is receiving exposures to the toxins many orders of magnitude lower than the smoker (and then zero after he leaves the room). The output of the cigarette is diluted into the air and since with toxins dose is everything; below a certain threshold there is no detectable effect of the toxin and somewhat lower than that, probably another threshold at which there is no effect, observable or not.
Moreover, the smoker likely smokes on a regular basis (often chain smoking) but with a few exceptions, exposure to second hand smoke is sporadic. That is why it is only possible to detect significant health effects on people who have lived or worked in close quarters with smokers for many years (and even then the effect is difficult to pick out which explains differences in conclusions and p-values between recent studies and even metastudies)
It’s exactly the same issue with wood burning. Sam makes his case, and then smugly conjures up a few straw man arguments without really considering the problem carefully. Most people don’t ‘chain’ burn wood. Even those who heat with it use it mostly in the cold months but most others will burn it once and a while, to cozy up the fire on a cold winter’s day. Most of the smoke goes up the chimney. Few people stick their faces right into the fire and breath deeply; the smoke (the little left over that doesn’t go up the chimney) is diluted orders of magnitude as it wafts through the air towards you. Who cares if it is 30 times as carcinogenetic, if its millions of fold diluted by the time it gets to you? The portion which makes it to other residences is diluted even more. BTW: Doesn’t it strike you as strange that he says 70 percent goes into other residences? The vast majority of space in our neighborhood is not taken up by houses, but filled with trees, lawns, roads and driveways. What is the magical magnetic attraction the smoke has to head straight into other residences and the concentrate- Sam does not explain – or do the smoke toxins just pass through briefly on their sojourn around the world which of course means even greater dilution into the atmosphere, and therefore even less likely to bother anyone?

Gus Van Horn said...


Thanks for raising these points, with which I agree, but simply didn't have time to pursue. I do disagree with Harris regarding both the magnitude of the danger of "second hand fireplace smoke" and what to do about it. (For example, someone could conceivably burn so much wood as to pose a nuisance to his neighbors, for which there are already legal remedies.)

Probably, I should have said something like, "I have no trouble entertaining the idea that wood fires are dangerous..."

IN any event, thanks for speaking up.


Snedcat said...

Part I

Yo, Gus, you write, I do disagree with Harris regarding both the magnitude of the danger of "second hand fireplace smoke" and what to do about it.

Yes, I had much the same reaction as you did when I read that screed of Harris's a few months ago, and many of the same objections to it as Steve detailed came immediately to mind. (That's what happens when you're scientifically trained and spend a lot of time reading anti-nuclear propaganda.) And you'll notice, as I did on rereading it, that Harris nowhere gives references for his claims; he just barrels through with "scientists this" and "scientists that." In place of connecting the dots with actual figures that address the crucial issues of dosage and exposure that Steve brought up, he coughs up (smoking pun intended) the usual anti-nuke and anti-cigarette scare mongering about "known carcinogens"--without addressing the question, known to be carcinogenic in what doses, and relative to what typical exposure?

And again, just as Steve pointed out, I suspect serious misrepresentation in his statement, "Research shows that nearly 70 percent of chimney smoke reenters nearby buildings." Oh, really? That would be suffocating to the neighbors of any wood burner, which is contrary, and flagrantly so, to everyday experience; perhaps it's a sloppy or even intentionally dishonest misstatement of something like, "measurable levels of smoke from a fireplace are found in 70% of the buildings within a certain distance from the fire," which is a believable but entirely different statement.

And the matter of dosage is especially important in statements like this: "In the developing world, the burning of solid fuel in the home is a genuine scourge, second only to poor sanitation as an environmental health risk. In 2000, the World Health Organization estimated that it caused nearly 2 million premature deaths each year..." Leaving aside the crucial questions of how this was "estimated" (a very vague word that could refer to anything from detailed epidemiological studies based on correlations between measured exposure levels and medical statistics, down to back of the envelope calculations or even utter and complete crap made up by some woozy-headed activist with ulterior motives looking for the next big pot to stir), there's the fact that (1) it's not at all clear that this is relevant to developed countries in the way Harris wants you to think it is, since often such fires are not built in well-ventilated fireplaces that we are familiar with, but are out in the middle of the main room or kitchen, and (2) "solid fuel" need not be wood--it could refer to coal or charcoal, or even animal dung.

Snedcat said...

Part II

Yes, you read that right, animal dung, which brings up a possibly relevant story. In Mongolia, the major fuel for fires on the steppes is dried livestock droppings, with certain types used as kindling, others as fuel. (Horse patties are dispreferred as fuel, for example, because they contain too much undigested grass that smokes up the joint quite terribly. Round droppings of sheep and goats, on the other hand, are I believe the usual kindling.) One of the major drives of the communist government in the late 1940s was to introduce metal stoves throughout the countryside in place of the open hearth fires that filled yurts with dense smoke. This was implemented quickly, and, according to health ministry figures, in the space of three years the infant mortality rate in the countryside decreased by two-thirds. (I think this was between 1946 and 1948, but I'm not sure.) So yes, I can see hearth fires being an important health problem in the developing world; I also see little relevance to our own fireplaces. But then again, he's not actually trying to make a dispassionate case about public health issues.

Thus, he writes: The case against burning wood is every bit as clear as the case against smoking cigarettes. Indeed, it is even clearer, because when you light a fire, you needlessly poison the air that everyone around you for miles must breathe. Even if you reject every intrusion of the “nanny state,” you should agree that the recreational burning of wood is unethical and should be illegal, especially in urban areas. Notice all the dubious laps in logic there, some of which Steve pointed out: (1) No, it's not at all clear as the case against smoking cigarettes per se; rather, it's as dubious as the case against second-hand smoke; (2) when you light a fire, you release smoke into the surroundings, yes, but absent any solid epidemiological figures it begs the question to say you "needlessly poison the air" around you; (3) it therefore does not follow in the least that lighting a wood fire is unethical, never mind legality. Quite frankly, it's a crap argument; you might say it's a lot of hot air filled with smoke from a horse-dung fire!

This is why this statement of Harris' strikes me as a complete failure of critical self-examination: I have discovered that when I make this case, even to highly intelligent and health-conscious men and women, a psychological truth quickly becomes as visible as a pair of clenched fists: They do not want to believe any of it. Most people I meet want to live in a world in which wood smoke is harmless. Indeed, they seem committed to living in such a world, regardless of the facts. No possibility of honest disagreement there, no sirree, just prejudiced and unthinking superstition chafing against enlightenment! Remember, this is based on their reactions to what Harris passes off as "rigorous research...conducted at dinner parties" (!!!); I seriously doubt that he fleshed out his figures and arguments there if he failed to give them in print here. I'm sure my objections occurred to most of his demurring listeners, but given the circumstances in which he crusaded on his high horse they were unlikely to go into the details or dwell much on the matter.

And if the case he presented against wood burning is his idea of air-tight argumentation, then I have no grounds to trust his evaluation of the psychological state of his interlocutors either, and no reason to conclude that opposition to his proposal to ban fireplaces is comparable psychologically to his atheist pronouncements. Indeed, it wouldn't surprise me in the least if he in fact holds his anti-wood-burning position on irrational grounds, for he certainly doesn't seem to think he needs to come up to the usual levels of evidence and proof accepted in science.

Snedcat said...

Note, by the way, that while Harris posts a link to a review article as recommended reading, it costs $43 to read unless you have a subscription. Granted, it's a bit unfair of me to say his argument's entirely citation-free, but it's unfair of him to make interested readers without subscriptions or access to a library pony up $43 just to check his work.

Snedcat said...

Yo, Gus, turns out there's a version of the article he cited available online here. It also turns out Harris possibly misrepresented the figures: "Exposures to biomass smoke are common in nearly half the households in the world that use wood, crop residues, or animal dung for cooking and heating...In more than a dozen studies each, two important diseases, chronic obstructive lung diseases and acute lower respiratory infections have been strongly associated with these household exposures, leading to an estimate by WHO of some 1.3 million premature deaths per year globally." (58). Not 2 million as Harris says, presumably in reference to the source cited in this study, but 1.3 million.

Also note the following in light of Harris's assertion that the scientific facts are thoroughly settled:

"Unfortunately, not enough is currently known to reliably distinguish the effects of different types of biomass smoke (e.g., smokes from wood versus crop wastes)" (57)


"Surprisingly few studies have been done in developed countries of the health impacts of woodsmoke, partly due to the difficulty of separating out the woodsmoke portion of the health effects of mixtures of particles in the environment." (58)

I'm not saying that wood smoke does not have adverse health effects, especially in poorly ventilated buildings in the developing world where it has to be used for heating and cooking, and in their surroundings; quite the contrary, the evidence suggests it does.

However, Harris overstates his case, appears to misrepresent his source by exaggerating the mortality figure it gives by more than 50% (!), and makes a dubious argument to attack all residential use everywhere in absolute terms ("There is no amount of wood smoke that is good to breathe," emphasis added)--no nuance, no attempt to compare wood smoke in such environments to other factors, no attempt to show that the typical concentrations due to recreational wood burning as opposed to other uses in North America exceed some threshold of measurable health effects, not even any attempt to show that recreational wood burning is in fact the major component of urban air pollution (as opposed to all residential use). Instead, Harris just equates wood burning with cigarette smoking!

The problem's not with people accepting or rejecting the case for ill health effects due to wood burning as such, it's with the attitude Harris displays, as if failure to buy his argument on first presentation is simply due to unreason and wishful thinking akin to religiosity. What the hell does he think practicing reason actually involves? Like we're supposed to change our minds with every slight breeze and enlist in the latest faddish campaign because of some propagandist haranguing us at a dinner party? Just because he's convinced himself, we're supposed to be convinced in lockstep with him, following our elite betters just on their say-so? What a pretentious authoritarian ass this Harris chap shows himself to be! No wonder he gets up people's noses and puts their backs up.

Gus Van Horn said...


Thanks for raking Harris over the coals on his anti-fire sentiment, the way he presented it, and especially his implication that there was no honest way to disagree with him.

Perhaps his very style of argument is also an example of what (actual) advocates of reason are up against.


Steve D said...

‘That's what happens when you're scientifically trained and spend a lot of time reading anti-nuclear propaganda.’

That sounds like me exactly. All I can add is that I have also read a lot of anti-GMO propaganda. In these blog posts they could almost just change the name of the REALLY REALLY BAD THING and then republish the same rant. (or program a computer to write it for them).

I glanced at the report with particular interest in the inhalation studies, all of which used unrealistic levels and timeframes which actually suggests the same sort of logic used against second hand smoke – maximize the exposure to find an effect, and then publicize it mercilessly. Though like Snedcat said, in this case the population and epidemiology studies cited were too hazy (pun alert) to make any clear generalizations. I did not go back and read the primary literature cited in the report but I suspect if I did I would find that Sam’s case is even weaker.

I’ve brought this issue up before. The problem with science these days is neither the data nor the methodology. It is in how the data is interpreted or in some cases the confidence level of the conclusion. The scientist needs to clear connect his conclusions with his data but often this gets lost in the publicity. I first noticed this problem years ago when I took a freshman psychology class in university. Many of major (good) experiments done in the history of psychology seemed to me to lead to only mundane or obvious conclusions but the professor would then fly way beyond the data and make conclusions which would make any biochemist or biologist cringe. Same issues with cosmology, quantum gravity theory, climatology etc. (notice I did not add sociology since I don’t believe it is actually a science, although physiology is, just not as presently practiced – and this is why chemistry and biology are the hard scientists today, not modern physics which in many of its manifestations hardly even ranks as a science at all anymore.)

Interpretation, interpretation, interpretation.

Gus Van Horn said...


You make me think of two almost-blogworthy recent examples of non-breakthroughs reported as if they were: an infant allegedly cured of HIV, and a diatomaceous "meteorite".

Laugh or cry: take your pick.


Snedcat said...

Steve D writes, "Interpretation, interpretation, interpretation." And that's the place where philosophy tends to come into the picture.

The problem with science these days is neither the data nor the methodology. It is in how the data is interpreted or in some cases the confidence level of the conclusion.

Quite so. I might add that this is glaringly obvious in the studies in which methods from the hard(er) sciences are applied to new fields in the social sciences or humanities, often with the implicit claim that without those methods the fields have not yielded actual solid knowledge. For example, there's a number of much ballyhooed studies in which various computational techniques from biology have been applied to problems of linguistic affiliation; such studies are readily published in Nature and Science without peer review by historical linguists and thus despite serious errors in data, sources, models, and interpretation; these journals are much less eager to publish replies.

For an egregious example, there's a series of posts at an online journal on historical geography by Asya Pereltsvaig and Martin Lewis about such a study by a multiple offender, Quentin Atkinson, on the location of the first speakers of Proto-Indo-European. There's a lot of dubiety in the study, so much so it took well over a dozen posts to treat it properly, including 5 posts devoted to 103 significant errors in the data on the geographical locations of the languages included in the study. Especially useful on the question of interpretation and of the models used are this one on a proper mathematical model for language spread, this one on the significant differences between linguistic and biological phylogenies, and this one on the basic failings of the model, with a representative quote:

But as long as the model rests on the untenable assumption that languages spread through a contagion-like process and diverge in speciation-like events, the result will still be of little value. Subsequent posts will examine how languages do spread and change. As we shall see, such linguistic processes are vastly more complex than the scenarios posited by the Science team. That does not mean that they cannot be mathematical modeled, only that any such efforts will have be much more involved than what we have seen thus far.

Snedcat said...

Part II:

But beyond that is the whole question of why such half-baked failures get published. Why are they so attractive to science journals? Why do they get puff pieces in The New York Times the day before they are officially published? (See Pereltsvaig's post on this; as she writes, "Overall, the debate is presented as a duel of name-calling and hand-waving between two camps of non-linguists: Quentin D. Atkinson and his colleagues on the one side and archeologist David W. Anthony on the other. No real arguments from either camp are addressed, making it appear that the two approaches are mutually exclusive, and that only one can ultimately be right. Such a caricature, however, [is] far from the truth, as several highly respected historical linguists, such as Don Ringe and April McMahon, take into account both the quantitative methods touted by Atkinson and the philological evidence explored by Anthony.")

Yes, there's a good deal of insularity in historical linguistics, as Lewis (an anthropologist) points out: Historical linguistics is currently in crisis not only because of unsubstantiated attacks or the failure of others to appreciate its intellectual achievements; it is also languishing because its practitioners have failed to meet the challenges that they face. All told, they have remained too insular and too comfortable with their own research paradigms. Emphasizing, like good scientists, the narrow acquisition of knowledge along established research fronts, few members of the guild have been willing to stand back and address the larger implications of their own work for the study of human pre-history (and history), let alone offer edification for a general audience. By the same token, few historical linguists have collaborated extensively with scholars in other disciplines.

But that does not change the fact that such studies are themselves examples of equally insular approaches, only extended imperialistically beyond the shores of their islands--a failure of wisdom rather than of vision, which is surely a worse failing.

Again quoting Lewis (for this captures the failings of the project quite well), "In the Science piece, the painstaking work of generations of historical linguists who have rigorously examined Indo-European origins and expansion is shrugged off as if it were of no account, even though the study itself rests entirely on the taken-for-granted work of linguists in establishing relations among languages based on words of common descent (cognates). In Wade’s New York Times article, contending accounts and lines of evidence are mentioned, but in a casual and slipshod manner. More problematic are the graphics offered by Bouckaert and company. The linguistic family trees generated by their model are clearly wrong, as we shall see in forthcoming posts. And on the website that accompanies the article, an animated map (“movie,” according to its creators) of Indo-European expansion is so error-riddled as to be amusing, and the conventional map on the same site is almost as bad. Mathematically intricate though it may be, the model employed by the authors nonetheless churns out demonstrably false information."

Sounds rather like the sliding scale of care in interpretation that we have noted here about Sam Harris, no?

Snedcat said...

Part III:

Lewis gets to the nub of the matter here:

"Failing the most basic tests of verification, the Bouckaert article typifies the kind of undue reductionism that sometimes gives scientific excursions into human history and behavior a bad name, based on the belief that a few key concepts linked to clever techniques can allow one to side-step complexity, promising mathematically elegant short-cuts to knowledge. While purporting to offer a truly scientific approach, Bouckaert et al. actually forward an example of scientism, or the inappropriate and overweening application of specific scientific techniques to problems that lie beyond their own purview." (Footnote omitted, bolding added)

And as we have discussed here before, scientism is a common enough error based in turn on deeper philosophical errors; the crudest form of it is that only mathematically secured knowledge is truly scientific knowledge, which suggests that the error is indebted to the analytic-synthetic dichotomy. Mathematics can be fruitfully applied in many fields, of course, but that does not mean that it is the only truly secure route to knowledge. More specifically though, is it the only route to scientific knowledge? That in turn depends on whether one takes the position that scientific knowledge at its best must give an account of the causes of what is being studied; and if this is not accepted, then there's no real objection in principle to applying any mathematical method in any realm of knowledge and pointing to statistically significant correlations as the final stopping point of scientific investigation, and thus in this instance dismissing all the knowledge discovered by two centuries of linguists relying on largely non-mathematical methods.

(I should add that Pereltsvaig, whom I have met, is primarily a syntactician, not a historical linguist as such, and Lewis is a historical geographer, not anthropologist as I stated earlier; he is the author of a pretty good book I recommended to Gus well over a decade ago, Green Delusions: An Environmentalist Critique of Radical Environmentalism. Their critiques are right on the money, though, even though certain subtle touches make it clear that they're not professional Indo-Europeanists.)

Snedcat said...

Part IV:

On a more positive methodological note, I should point out how mathematics could be used for such studies; since I do a good deal of quantitative work in linguistics and am trained as a historical linguist, I can contribute something useful on that count.

One might think that since languages change in such a way that it prima facie resembles biological speciation, then the mathematical tools developed for work in cladistics and genetics could be immediately applied to languages. However, this is not so: Languages do not show the difference between ontogeny and phylogeny that organisms do; language changes in Lamarckian rather than Darwinian fashion. There's no "genetic code" to a language that might mutate but that is otherwise transmitted to later generations of speakers unchanged by environmental factors (and in this case extralinguistic social factors in particular)--unchanged apart from selective pressures exerted through their effects on survival and reproductive success, that is.

In fact, although it's easy enough to see idiolects (the individual varieties of a language spoken by each of its speakers) as comparable to the individuals in a population of a species, once you think about it, it's hard to see what feature or level of language would be comparable to the genetic code. Words? Sounds? But people pick up new words and lose old ones all the time form their neighbors, and their pronunciation of the sounds in their speech changes over their lifespan--indeed, one of the major breakthroughs in historical linguistics in the past 50 years has been the ability to measure change in progress quantitatively, both in words used and sounds produced, and these changes spread through a population from speaker to speaker on the time span of the order of a decade, quite unlike genetic changes in a biological population, in which changes spread by differential rates of propoagation in later generations. And on this count I refer you to the posting I linked to before about more complex mathematical models of the spread of language change.

Moreover, such changes are channelled (preferentially spread) through social networks, spread at widely different rates depending on the ages of speakers (early teen to early twenties being most apt to adopt changes, at least in industrial societies in which your daily contacts and social networks mostly involve people of your own generation), and often are semiconscious--they get adopted more readily if they are seen as prestigious (though "prestigious" is a very broad cover term that includes a wide variety of different types of prestige).

To the extent a change is fully unconscious, it spreads with greater regularity throughout a speech community; this is especially true of minor changes of pronunciation without any connection to independent meaning. In the case of changes in pronunciation, the less arbitrary the connection of sound and sense, the more readily speakers will become conscious of the difference and either adopt it or react against it (for reasons of prestige, group membership, and the like). Such changes that are due to essentially mechanical factors in the speech act are the prototypical examples of "sound changes"--the name is misleading, for they are not all changes in the the sounds of a language, but only the changes in sounds due to the effects of neighboring sounds--concretely neighboring or adjacent in the stream of speech itself. Pure sound changes spread exceptionlessly within a given speech community (usually in a decade or less, depending on the size of the community), though they can be reversed in certain words or smaller meaningful units (morphemes) to the extent there's an association of meaning involved; hence the usual tool of finding the exceptionless sound changes first, then striving to explain the exceptions on the basis of common patterns in languages.

Snedcat said...

Part V:

That is, to the extent the change in pronunciation affects a sound that has some meaning associated with it, such a change can be rejected or modified; in technical terms, these are changes by analogy: A variation in vowel length might become associated with a difference in tense, for example, and get extended to other verbs not showing the variation (a common occurrence in Middle English, for example); or a change in pronunciation might affect one variant of a morpheme (an ending with two or more distinct forms depending on the root), such that speakers then correct the variant to an analogous form (this is especially common if the ending is in a paradigm, in which case it is called analogical levelling and affects a big slew of forms in the language, but it might also affect a less common form that is then changed on the model of a similar-sounding more common form, which is called "four-part analogy"). Most of the time even analogical changes are largely unconscious and only act over centuries, hence the importance and the usefulness of nailing down the sound changes first, since they account for probably 90% or more of the changes you'll find and constitute the background against which most analogical changes occur.

Moreover, sound changes are not all equally probable by any stretch of the imagination. The change from tea to chee is very common and quite probable; the change in the opposite direction is extremely unlikely. Part of the training of a historical linguist is to become quite familiar on a largely implicit basis with the relative likelihoods of sound changes through study of the histories of a wide range of languages and language families, especially of attested sound changes (as in the history of the Romance languages, etc.); while there is some good, very interesting research on the physical and psychological factors (ease of articulation and perception) contributing to various classes of sound change (the linguist John Ohala is closely associated with some of it), there's no likelihood of a fully explicit, mechanical procedure for predicting or retrodicting sound changes since (1) there are no absolute optima involved, most basically because articulatory and perceptual factors are by nature usually opposed, (2) for all but contemporary forms of the language there's no way of determining the necessary details of pronunciation (relative timings, strengths of production, degree of coarticulation--the mutual influences of sounds upon each other when spoken in succession due to having to move the tongue and jaw from one position to another), and (3) sound changes spread or die out for extra-linguistic social reasons. Instead, it's a problem of historical interpretation--how does one use observable facts to infer the sequence of past events that led to them, and how secure can one take such inferences to be?

Snedcat said...

Finally, Part VI:

And such inferences are complicated by the fact that in and of itself and in isolation from the rest of the facts of the languages involved it is usually impossible to determine if a given resemblance between two words in different languages is due to common descent or borrowing; this can only be determined (to take a very rough view of things) through the historical record (if we're lucky enough to have such), the system of sounds and sequence of sound changes over time inferred for the two languages on the basis of the facts of the languages as a whole, and factors of and patterns in meaning of those and related words, here ranked from most to least secure. (There's also the possibility of coincidence and of non-arbitrary connections of sound and sense: Onomatopoeia and much baby talk, for example.) Certainly there's no factor like the genetic code to restrict a given resemblance to necessarily falling in one category or the other.

This is one place where computational methods are quite well suited to historical linguistics, churning and processing reams upon reams of data, if you're lucky to have them. Simply glomming onto gleaned and cherry-picked data like some illiterate Young Turk and shoving it into off-the-shelf cladistics software, on the other hand, is an exercise in arrogant futility--but it does seem to be a choice route to fame and enhanced funding in some jurisdictions. And as the example of the paper on Indo-European whose critique I linked to so thoroughly shows, you'll also be very likely to find a Wade or a Harris eager to trumpet your findings to the four corners and the skies without expressing any of the reservations you yourself might be too honest to omit.