Sunday, May 29, 2016

When is absence of evidence evidence of absence?


“Absence of evidence is not evidence of absence.” If this isn’t the most-quoted maxim in cryptozoology, I don’t know what is.  But let’s think about it a bit.

It’s not always true and not always false.  There’s a sliding scale with some subjective judgments involved.  I’m going to make a qualification here: I’m defining AOEINEA as evidence in addition to, or in lieu of, eyewitness reports.  That’s not always true and not always fair, but interested scientists may accept some eyewitness reports and dismiss others depending on the perceived probability of the animal (if you say you saw a new rodent, maybe you did: if you say you saw a fire-breathing dragon, you didn't, because they're impossible, and AOEINEA is not going to help you.) So let’s talk about other types of evidence. These are, of course, my opinions: if you don't like them, feel free to open fire.  

We will take a few hypotheticals, going from one end of the scale to the other. I am NOT arguing for the existence or nonexistence of particular animals: I’m examining whether the AOEINEA paradigm works for them or against them. There is some point, with any animal, where absence of evidence does become evidence of absence, but it’s not as clear-cut as all of us would like.
If I report seeing an odd-looking fish in the ocean, it can hardly be argued it doesn’t exist because we don’t have a specimen (specimens being doubly important in the ocean, where a creature doesn’t leave footprints or half-eaten meals or anything else behind it.)  A biologist will usually say, correctly, “We need a specimen to classify it,” but she has no a priori reason to argue it does not exist unless there is some major incongruity, such as me saying it has four pairs of fins.  Some of William Beebe’s reported fishes are dismissed because they are fairly large but have never been seen again: I would argue, though, that AOEINEA still applies because of the sheer size of the relevant habitat.
If I report that fish from a lake where humans have been studying and fishing for a long time, the AOEINEA paradigm is harder to apply, but not impossible: as in the shoal bass of the United States, it may just be that no one has looked hard enough at the evidence. If I think I saw a new kind of trout, ichthyologists may or may not consider it plausible.
If I report a small bird that doesn’t match any known type, absence of evidence is not to be held against it if I see it in a thinly populated region. Birds are elusive: they are easy to see but not to get specimens of unless you find a nest. They also have distinctive cries or songs, which are important evidence in ornithology.  Birds can even hover under the radar where you’d think the evidence should be there – my favorite example is the Brazilian ovenbird that nested yards from a major highway. 
Small mammals are pretty elusive, too. It’s been decades since a new species was described from the wild in the United States, but if I approach a mammologist with a good description of an odd-looking rodent, he’s likely to take some interest.  AOEINEA applies.

A new article from Audubon magazine describes the arduous search for the Cuban population of the ivory-billed woodpecker by making the bird's double-knocks and waiting for a reply thus:   "To get one to do so on this trip in a territory this large, he [ornithologist Martjan Lammertink] conceded to the photographer, would be very lucky. To not get one proves nothing."

If I report a very big fish from a known lake, it gets chancy.  I have argued that AOEINEA applies well to Lake Iliamna because 1) the eyewitness reports are good and pretty consistent, and 2) there is a known species in adjoining waters (white sturgeon) that has an unimpeded path to the lake if it has chosen to make the trip. 
If I report the most spectacular type of new species - a large land mammal -  it’s somewhat type-dependent, and somewhat location-dependent. If I see a new tapir in the Amazon, science will want a specimen but won’t dismiss it out of hand: there’s a lot of virgin forest left, and new species come out of it all the time.  If I report a small ground sloth, that’s harder to swallow, but I’ll stand by AOEINEA as long as there’s adequate food, an ecological niche where it won’t be hopelessly out-competed (as the mainland thylacine of Australia seems to have been by the numerous, fast-breeding dingo) and  room for a population to hide away from humans.

Now let’s graduate to lake “monsters.” If witnesses report a big animal in a lake that may or may not be a known species, we can point to our friends the sturgeon, gar, wels, or other large freshwater fish and argue AOEINEA if we’re in or near their reported range. Again, Lake Iliamna is a good example, If witnesses get a good look at such an animal and are certain it can’t be a known species, then things get complicated. The AOEINEA paradigm does not cease to apply, but it’s not as firm:  there should be fossil evidence of large creatures, evidence of predation, etc. even if there are no catches or carcasses.  Granted that the fossil record is very much incomplete, you can argue AOEINEA, but there’s a caveat. If the creature we postulate has a very sparse fossil record, that should be ok, but if it has a rich fossil record that stops dead millions of years back, our creature is in trouble.  It can happen (cf the coelacanth) but is less likely.  Complete absence of fossils, paradoxically, may be easier to swallow: that sounds highly illogical, but consider that we have no fossil evidence for the saola, just the live animal wandering around Vietnam and Laos, and until a few years ago there was no evidence for chimps despite many thousands of them being alive and probably millions having died in years past.  If there is a fossil record, we know the animal lived under conditions where it coulod fossilize and in a place where fossils could be found, so the question of why they should stop is a very troubling one.  

Now habitat comes back into play. If we have good sightings of a new ape in Sumatra, which we know is good ape habitat and still has large wild areas, AOEINEA stands up pretty well.  If we have them in the forests of the Pacific Northwest, most scientists would say that 50+ years of looking indicates absence, although AOEINEA isn’t quite dead as long as there is plenty of untrodden forest. If we have a similar sighting from New York’s Central Park, AOEINEA doesn’t work: while a few animals could hide briefly in the park, the idea of a long-term population leaving no hard evidence doesn’t work, and even the most romantic-minded cryptozoologist is not going to argue differently.

Let’s go back again to lakes.  If a creature of crocodile size or larger is reported from a small pond, and searchers cannot find it, then absence of evidence pretty solidly does mean evidence of (current) absence.   If it’s in a big lake, then AOEINEA applies until the lake is well searched. But what is “well?” Different definitions can be offered.  To take the biggest example, Loch Ness, I would argue AOEINEA applied in the early days, certainly up into the 1970s. Then things got problematic. In 2016, I think AOEINEA is dead for Loch Ness: if it existed, we would have better evidence from a large lake searched many times with different technologies.  We wouldn’t be arguing over photographs and sonar tracings 40 or 50 years old, because newer and better evidence would have superseded them.
Jumping over to Lake Champlain, I think the lack of evidence is a killer because of the habitat: the lake freezes over, yet no animals come out, knock big air holes in the ice, etc.  If you say it hibernates, unlike every other large marine animal known to science, this may be worth thinking about but becomes special pleading pretty quickly if no one can find nest sites, caves, overhangs with air pockets, etc. 
Of other famous cryptids, the yeti has much less evidence than sasquatch but a better plea for AOEINEA because of the remoteness of its habitat. “Sea serpents” can find some shelter under the postulate: while actual serpents are ruled out, we’re still finding sharks and beaked whales and the roiling habitat of 321 million cubic miles could still shelter a huge eel or elongated fish or -well, something.  Thunderbirds don’t get to use AOEINEA because they are extremely conspicuous in flight but seemingly invisible to tens of thousands of birdwatchers and a hundred million plus cameras. The evidence, not just photos but things such as huge nests and eggs, should be better. Likewise for primates in populated areas of North America, which also have to overcome, not only lack of fossils for their own species (which might be ok), but lack of any fossil apes anywhere in the New World.  Mokele-mbembe has passed the AOEINEA point with searches of its habitat combined with its size and visibility and its high degree of improbability – the habitat is not a “lost world” unchanged since the Mesozoic, and the fossil record for the dinosauria stops dead at (or microscopically close to) to the K-Pg boundary.



The Cuban population of the ivory-billed woodpecker: lost for decades, found in 1986, now back on the missing list.

On balance, AOEINEA has some utility but is not universally true.  (By the way, I reject, for cryptozoology, that other maxim, “Extraordinary claims require extraordinary evidence” – the evidence required to publish a scientific description of a sasquatch is exactly that required for a new mouse, a type specimen.)  You could easily create a list of cases (woodland bison, Itombwe owl, etc.) where AOEINEA was proven to be a good approach and many more where it got us nowhere.  In so many cases, this is still a judgment call and people are going to differ for a long time yet.  If you think that’s a wishy-washy conclusion you’re right: I’m not very satisfied with it myself. But that’s the real world for you.  

2 comments:

jamesrav said...

I believe you properly 'kill off' the major cryptids, with of course Sasquatch properly left off the list since it's still very viable. And really, do any of the others matter when compared to what would be the closest human relative? A 4-foot long salamander in the Trinity Lakes ... ho hum compared to a 7 foot tall bipedal 'apeman' that has eluded positive ID for 100 years. Getting away from your topic a bit, but my own conclusion about Sasquatch is strictly based on PG, without that it's a no-go for me. Footprints? eyewitnesses? they hold no sway and here's why: as a kid in the early 70's, the overwhelming 'strange but true' topic was of course UFO's. And that totally puzzled me - why was there any doubt? Books (dozens of them) had been written, including eyewitness accounts from AMERICANS (ya know, people just like 'you and me') who claimed to watch objects land and various types of aliens emerge. End of story. I mean cmon, how do you misinterpret something like that? Forty years later, older and wiser ...

Matt Bille said...

James,
I agree it makes sasquatch very difficult to advocate for. The Trinity alps salamander always interested me because, while no one has found one, the existence of similar-sized animals elsewhere means it's possible.
Hmm, UFOs... we have no doubt people are seeing and reporting unusual things, but as Hynek wrote, we don't have UFOs to study, only UFO reports, which is not very satisfying. (Although the term UFO is a bit problematic: I think it was Jenny Randles who noted that a thing sighted in the sky may not be technically "flying" or an "object." ) The landing/alien reports are like sasquatch reports: plenty of reports, plenty of seemingly normal people insisting they saw something extraordinary, but not one fragment of material. I do think the late Philip Klass was onto something when he suggested a larger, longer-lived cousin to ball lightning existed. I'm quite convinced the US government has no proof of aliens or contact with aliens: such events would shape the classified portions of our military and intelligence budgets, and yet a massive leak of our most sensitive documents included not a single word. So I think there's a natural phenomenon involved, but that's as far as it goes.