Mark Rosenfelder asks "How likely are chance resemblances between languages?" and gives his answer (link).

I'm not up on statistical analysis, it has been a long time since university. And unlike journal articles he doesn't have a conclusion section at the end. If anyone understood all those graphs could you translate for me please?

Way outside my realm of understanding, but various parts were
very enlightening. Thanks.

If anyone understood all those graphs could you translate for me please?

There was only one graph, the normal distribution one, and I was not sure what his point was, other than the distribution of probabilities is a normal one. It's been a long time since I did statistics, too. Perhaps one of the other mathematically inclined can vet his reasoning. I mainly concentrated on his explanations of how the mass comparison "method" works.

1. The lack of consistent phonological correspondence between languages (say as between Latin and French or Italian, for example).

2. The phonological and semantic leeway in the mass comparison "method".

His conclusion I took to be that the kinds of mass comparisons that some people find convincing could be no more than coincidence.

Done my best to read the complete article, but when cifers are getting involved, I just loose track. Though I can grasp some of the contents. We've had comparisons before, of which you said they were accidental resemblances. I looked twice in surprise at today's word "ort".

ort
"remains of food left from a meal," c.1440, cognate with early Du. ooraete, Low Ger. ort, from or-, privative prefix, + etan "to eat."

Ort is a word I never heard of before which is exactly written like German "Ort" meaning "place". Would you know where the German word Ort comes from? It really can be so confusing.

German "Ort" meaning "place". Would you know where the German word Ort comes from?

I did not, but I took a look at Kluge and Grimm's dictionary. In Middle High German, ort meant 'point (of a weapon)'. It developed a secondary meaning of 'corner; angle', and from that developed the meaning of 'place' in modern German. It had the point of a weapon meaning in Old Saxon (Dutch) and old English ord. (This latter is not related to ordnance (var., ordinance) which is ultimately from Latin, via French, ordo 'order'.

What always intrigues me is where you say : "it developed". It developed from 'point' via 'corner' and 'angle' to 'place'.
But how and why we'll never know. How does the point of a weapon turn into place. It is strange juggling, language. Through history.

"it developed" It developed from 'point' via 'corner' and 'angle' to 'place'. But how and why we'll never know. How does the point of a weapon turn into place.

Well, I condensed my rather quick skimming of the long entry in Grimm's dictionary. It's online and you can follow the development here (link).

[Addendum: another odd development. In Old Icelandic, oddi meant both 'point' and 'three' or 'odd'. My Dutch etymological dictionary shows the same situation holds in Dutch: "oord ... "punt, kant, rand, spits van een wapen, hoek, vierde deel ... stuk land, tijdstip, begin".]

Perhaps through the pointing of a weapon, or some other angular/pointed instrument at a map to denote a location?

Perhaps through the pointing of a weapon, or some other angular/pointed instrument at a map to denote a location?

We use point to mean a place also in English. (cf. "We stopped at a point file miles short of our final destination", "The train stops at Chicago and all poionts west".) Looking more closely at the Dutch definition of oord and the Grimm's entry on ort, I see that 'point', 'edge', 'corner', 'border', starting-point', 'end-point', and 'place'. It has a logic of successive meanings to me that kind of seemed obvious, but mustn't be to others. English place is from Greek, via French and latin, and referred originally to a πλατεια οδος (plateia hodos) 'broad street'; cognate with English flat.

Rather than coming to a conclusion, he seems to be starting off making a point, then using all those calculations to "prove" it.
Essentially he seems to be saying that people, particularly those with no linguistic training, tend to see cognates and patterns in sounds and meaning(s) between languages, when statistically they are likely to be due to chance.

I can't argue either way, certainly. There are the physical limitations of the human speech structures--which is more of an argument for coincidence. But it seems to me that his parameters are rather arbitrary--though frankly I don't know how else he could have established them! But I can't help thinking that his numbers would be more accurate if: a.) we knew how things were pronounced in ancient times; b.) they took into account b.1.) travel and b.2.) regional accents and dialects.

Yes, I know that if he/we had that information, this whole question would be moot. But my point is that, no matter how intricately he delved into the numbers of known "phonetic leeway", I think there are too many unknowns for these numbers to be considered accurate.

It just seems to be one of those unanswerable questions; in any example, is the similarity pure coincidence or is there a piece of history that we're not aware of?

It just seems to be one of those unanswerable questions - I cannot but emphatically agree!
The question seems to me akin to asking: how likely are chance resemblances between clouds..

Mama and papa or baba are the first sounds a baby makes so in almost every language this becomes mother and father. This is not evidence of a proto language but coincidence. Are there others that anyone can think of?

In Georgian mama means 'father' and deda means 'mother'. See this excellent little paper.

WARNING: pdf

Thanks, that was interesting.

Very, very interesting. Thanks.