On 4/4/23 7:36 PM, Peter Nyikos wrote:
> On Monday, April 3, 2023 at 8:25:17 PM UTC-4, John Harshman wrote:
>> On 4/3/23 2:20 PM, Peter Nyikos wrote:
>>> On Saturday, April 1, 2023 at 7:29:22 PM UTC-4, John Harshman wrote:
>>>> On 4/1/23 7:53 AM, Popping Mad wrote:
>>>>> On 3/31/23 22:03, John Harshman wrote:
>>>>>> The immediate subject isn't computational math;
>>>>>
>>>>> Actually, that is EXACTLY the subject.
>>>
>>>> How so?
>>>
>>> You are missing Ruben's point, just as you did way back in 2016.
>
>> What is Ruben's point, then?
>
> It would be best if he were to state it in his own words. However, I can say this much:
> if "patristic distance" does not involve computational math, then you are wrong
> about it being the same thing as "path metric."

Of course it involves computational math. It's the distance between two
terminal taxa, measured along the branches of the tree, as opposed to a
simple count of pairwise differences.

>>> See my comments on a thread Ruben ("Popping Mad") began at 6:11 am EDT today:
>>>
>>> https://groups.google.com/g/sci.bio.paleontology/c/rX_v-fcwq7s/m/3Y7nZwYrAwAJ
>>> Re: Phylogenetic Trees: The What and The Why
>>>
>>> [EXCERPT:]
>>> For this first post, I am skipping over how phylogenetic diversity is calculated,
>>> and moving to an issue that has caused untold confusion down through the decades,
>>> both in s.b.p. and in the big outside world: how closely related are two species?
>>>
>>> This is quantified in the article by using the best weights we can assign to each edge in the tree,
>>> and adding together the ones on the unique path through the tree from species A to species B.
>>> This is called "the path metric" between A and B.
>>>
>>> The article avoids defining "related" [specifically, "more closely related"]
>>> because of the way the systematists who dominate systematics define it.
>>> Their method cannot be quantified because of its rudimentary nature.
>>> And it is totally at odds with the quantification I gave above.
>>>
>>> The dominant definition is the analogue of saying,
>>> "Mitochondrial Eve is more closely related to everyone on earth today than she was
>>> to anyone alive before she had children, including her parents and her siblings, if any."
>>>
>>> On the other hand, look at the sentence "Species A is more closely related to species B
>>> than it is to species C because the path metric from A to C is greater than the one from A to B."
>>> This corresponds closely to the way we, including genealogists, use "more closely related" in everyday life.
>>> [END OF EXCERPT]
>
>> What you call "path distance" is called "patristic distance" in
>> phylogenetics.
>
> Are you sure these designate the same thing? Ruben linked the following review
> article, which I was using for "path metric," and it is quite computational.
>
> https://people.math.wisc.edu/~roch/research_files/review-steel-ams.pdf
>
> It's a review of the following book (for want of a better word):
> Phylogeny—discrete and random processes in evolution1 by Mike Steel, CBMS-NSF Regional Conference Series in Applied Mathematics, 89, Society for Industrial and Applied Mathematics (SIAM), Philadelphia, PA, 2016. xvi+293 pp
>
> By the way, the review does not use the word "patristic" anywhere:

Yes, Steel, in all his publications, tends to use the language of
mathematics rather than that of phylogenetics. He's a mathematician, not
a biologist. He consistently uses "leaves", for example. But the
meanings correspond to phylogenetics terms.

>> But I wouldn't call that a measure of relatedness,
>
> That's because you are locked in to the definition of "relatledness" which triumphed
> during the cladist wars. The famous cow - lungfish -salmon cladist victory
> could have been avoided if someone who was well clued into vertebrate paleontology
> and its evolutionary trees had been given a hearing during that debate.

Why would we want to avoid it. Are you saying that salmon are more
closely related to lungfish than to cows? Would you also say that hippos
are more closely related to cows than to whales? Where does this end?

>> and it certainly doesn't fit the ordinary meaning, especially if evolutionary
>> rates differ among lineages.
>
> The ordinary meaning accommodates cases where a sequence of four successive generations
> starting with one couple occurs in the same amount of time as a sequence of two generations
> comes from another couple.

> Perhaps you forgot, also, that one extended family could remain at the same level wrt class or income or influence
> for many generations while another improves dramatically and while a third goes to pot.
>
> Different rates of evolution, y'see.

What does class or income have to do with relationships?

>> The cladistic definition could be
>> quantified on a time-calibrated tree, i.e. one on which the ages of the
>> ancestral nodes are estimated.
>
> This would not work very well in the two kinds of everyday examples I gave.
> Time calibration would not shed much light on the social dynamics involved.

What do social dynamics have to do with it?

>> But since in general this information is
>> unavailable (or at least unreliable), we content ourselves with only
>> relative ages, such that ancestral nodes are older than descendant nodes.
>
> Looks like you are using a concept that uses only extant organisms
> and ignores fossils.

It's certainly simpler for extant organisms than for fossils. But your
concept doesn't work very well for extant organisms. And most known
species are extant, so shouldn't they be accommodated?

>> I reject the idea that evolutionary distances should be equated with
>> closeness of relationship.
>
> But the one that depends only on the topology should be, in your opinion. Why,
> What, besides your loyalty to the dominant group of systematics, accounts for this?

Two main reasons: 1) Topology is objectively determinable, while
distance depends entirely on the data you choose to measure, and
different data are not necessarily correlated; thus relationships under
your definition can change radically depending on how and what you
choose to measure; 2) differences in evolutionary rates (for whatever
data you use) can lead to odd results, even those you would, I hope,
think were odd. Though I may be wrong. Would you think that magpie geese
are more closely related to screamers than to ducks? Genetic distances
would make that claim under your definition.

> Why has not anyone seen that fossils alone gave birth to a theory about the formation of the earth, that without them, no one would have ever dreamed that there were successive epochs in the formation of the globe.
> —Georges Cuvier

Of course, Cuvier was writing before Hutton, Lyell, or, especially,
radiometric dating.