Antoine de Saint-Exupéry is narrowly connected

+11 votes

When I saw that I was 39 degrees away from Antoine de Saint-Exupéry in today's WikiTree Family News I was a bit surprised (most of my connections are in the mid 20s). So I decided to give a look and actually Antoine is on average 43.21 degrees away from the average WikiTree profile (on the main tree).

I have recently been doing some graph analysis of the Wikitree connection graph. One of the things I've looked into most is how "central" a person is to the Wikitree main global tree. I measure this by the average (mean) distance to all other people in the tree by the same connections used in the connection finder (child/parent, spouse, sibling).

To give a bit of context here, here are the average distances to some other famous folks:

Person Average distance
Antoine de Saint-Exupéry 43.2145514495
Queen Victoria 22.4378277882
Chet Atkins 23.6506553683
Alexander Pushkin 26.3612622736

And in fact, digging in a little more, I see that Antoine is very narrowly connected to the main Tree. There are series of 29 connections from Antoine de Saint-Exupéry to Pauline Bonaparte where if any one of them were cut, Antoine would not be connected to the main tree, almost all connections go through these people.

To see this, look at the average distance from each of these people below. It starts with Antoine and each subsequent line is the person connected to the last person who has the smallest average distance. As you can see, that average distance goes down by almost 1 each step until Bonaparte-11 (Pauline Bonaparte) indicating that each person before that had almost 100% of connections going through them:

Saint-Exupéry-1 43.2145514495

Saint-Exupéry-2 42.2145532948

Saint-Exupéry-5 41.2145558199

Saint-Exupéry-6 40.2145589278

Quatrebarbes-2 39.2145624241

Quatrebarbes-3 38.2145662118

Quatrebarbes-4 37.214570388

Quatrebarbes-5 36.2145750498

Quatrebarbes-8 35.2145810712

Quatrebarbes-10 34.2145882581

Quatrebarbes-11 33.2145963191

Chevallier-Chantepie-3 32.2146052542

De Gaulle-8 31.2146143835

De Gaulle-1 30.2146303113

De Gaulle-2 29.2146509008

De Gaulle-9 28.2146726558

Bidauld-2 27.2146950906

Lami-14 26.2147178167

Lami-15 25.2147408343

Borghese-3 24.2147645316

Borghese-4 23.2155929685

Bonaparte-11 22.2159523139

Bonaparte-30 21.6475706308

Murat-9 21.188609139

Willis-5459 20.476588161

Willis-1442 19.6930025737

Willis-1050 19.0155785947

Willis-1049 18.7825173603

Lewis-3843 18.4671508765

Lewis-2595 18.4211808867

WikiTree profile: Antoine de Saint-Exupéry
in The Tree House by Shawn Ligocki G2G6 Mach 1 (16.6k points)
edited by Shawn Ligocki
I'm sorry, I''m not sure I understand. Is this a criticism? Antoine de Saint-Exupéry is French, and therefore far away from the core population of WikiTree profiles. I'm very grateful he was selected as Profile of The Week however.

You're very welcome to help finding better connection paths for his profile. It was quite hard work finding the one path that exists, I can tell you.
I don't read Shawn's post as a criticism at all.

He is just interested in the statistics of who is solidly connected to the main tree - by being connected to lots of other people - and who is not.

That's all.

One of my notables Helene Hanff (author of the book 84 Charing Cross road) has only one very distant connection to the tree - and if that connection should be broken then she loses her connection to the tree.

Helene was of Jewish origin and it took a while for that connection to be found.

Karen Carpenter is another Notable who has just one distant and tenuous connection to the global tree. If that connection should be broken, then she would NO LONGER be connected at all.

Karen Carpenter was an all-american girl and it took more than 6 months to finally get that connection. You would that someone whose family lived in Baltimore for 6 generations would have an easy connection, but that did not happen.

This is just a FYI thing, Isabelle, that's all. For those who are interested. Such as me.
The connection from Saint-Ex to the de Gaulle family is not particularly weak. Yes, it's only a single chain, but all the profiles in the chain represent dead people (so they won't be deleted), and I'm satisfied with the sources linking each to the other (a very pretentious statement, I know). So IMO though there is only one chain, the risk of breaking it is minimal.

From there, the de Gaulle family has three connections to the main tree (that I know of) including the Bonaparte one which IS tenuous, as the evidence connecting Lami and Borghese is weak (and if I were to create those profiles now, I would not link them).
Sorry for being a bit unclear. I did not mean to make any criticism! I also don't mean to imply that any of the connections were dubious. I have changed the wording from "tenuously" -> "narrowly". As Robynne said, I only brought it up out of general curiousity in the structure of the main Tree. Although Antoine de Saint-Exupéry is French (and it appears the majority of WikiTree is American), I imagine a lot of people in WikiTree are connected to him more closely, we just haven't found all those other connections yet.

That said, thanks for all the work connecting him this much! One thing that interests me in this line of research is evaluating the "value" of various connections in bringing people closer to the "center" of WikiTree. I think it would be nice to be able to have a way to measure that you had added connections which bring Antoine de Saint-Exupéry closer to the center :)
Thanks to both of you for the clarification.
I also tend to connect to folks in the mid-20s so I'm working to shorten the paths to Saint-Exupéry. It doesn't mean we don't love the work that has been done - we just want to do more work!

There's worse than Saint-Exupéry though. Andrés Avelino Caceres, President of Peru, is truly sitting at the end of the connection chain as father of an ex-wife of Consuelo de Saint-Exupéry's second husband. Without Saint-Exupéry, he would not be connected at all. Now if someone could do something for him.... 

Robynne, LOVE the book "84 Charing Cross", haven't thought of it in years, ten times better than the movie. I urge everyone to read it.

It is not just the French connection.  Stephen Hawking was another one that was remotely connected (to me anyway). 

Or rather, it's not just Antoine de Saint-Exupéry. You should look at René Descartes's connection (Descartes-3). Or Denis Diderot (Diderot-1). Or Braille-1, or Davy_de_la_Pailleterie-3... most of the French  notables I connected, anyway.
How do you figure out what someone's only connection is to the global wikitree ?

What is Karen Carpenter's only connection to the global wikitree ?

I can search for more connections for her.
I spent months working on the Carpenter family tree.

The only connection that I am aware of is through the wife of a second cousin from the Tatum Family - which is the Maternal line for Richard and Karen.
I found another connection for Karen Carpenter through her Guyton ancestor. I am currently looking for sources and entering them as I go. I just initiated a Guyton merge with different degrees, so that will mean one more link through a different Guyton descendant's marriage, I think.
Carpenter's Guyton is now connected back to the 1600s and the global wikitree.

I can also add ancestors for Abel Briggs ... however I think there is questionable link there. Able Briggs has a son born in 1847 called Wellington which conflicts with what wikitree has as George Delvan born that year (who's parents I don't know).

update: I found the 1855 state census and it looks like George Delvan was born in 1845, I will see how far I can go back with Able Briggs.
So for Karen Carpenter I have taken some of her lines back to the 1600s giving her more global tree connections for each ... ...

on the above page I did Guyton and Briggs (& Vaughan) and for Ingley, I did the Wood line.

Note: Some profiles won't show connected until tomorrow's update.

5 Answers

+10 votes
For a numbers freak like me, this was a great post. Thanks, Shawn.
by Pip Sheppard G2G Astronaut (2.2m points)
+5 votes
Regarding the average distance of French profiles, it's a bit difficult to generalize.  Those profiles that connect within the continental European population tend to be much more distant from the center than others.  However, most of the French notables I've connected have gone through the US and Canada, and those wouldn't be much more distant than the typical WikiTree profile.  I checked the connections of those profiles (I have a list on my own profile) to featured WikiTreer Pip Sheppard, and the numbers are in the typical 20s.  Of course, not all French profiles are easily connectable to North America, and those that can't will remain more distant for the foreseeable future.

As for determining another path to connect Saint-Ex, the Canadian and French Taschereau families are different branches of the same family of minor nobility from Touraine.  It may be possible to find the common ancestor, though I've already determined that it would be prior to 1600.  Another possibility is looking at the Bellegarde family in Tours, which marries into both sides of the Taschereau family.
by Greg Lavoie G2G6 Pilot (315k points)
edited by Greg Lavoie

Those who are interested in finding out more information about the Taschereau family may find the book La famille Taschereau to be of interest.  It's apparently in the public domain and available to view and download at the link here.

+5 votes
I checked the number of degrees for many of the French notables on my connections list (visible on my profile) to this week's featured member William Foster. There is a small group at 26 degrees, but everyone else is more than 30 degrees. The record is Louis Braille (Braille-1) who is 46 degrees from William Foster.

The only exception is Claude Monet, who is only 20 degrees away from William, and that's because his wife's daughter happened to marry into a patrician Philadelphian family.

This exactly confirms what Greg says above. That, or I really suck at connecting ;-)
by Isabelle Martin G2G6 Pilot (452k points)
+8 votes
Unearthing this last year thread, sorry, I was not here at the time :-)

I'm currently adding some French more or less notable profiles similar to "Saint-Ex": descendants of Euro aristo families between 1850 and 1950. So far they are connected to the tree only through my wife and me, and an obscure ancestor of mine. Very long and tortuous path.

I know all those folks have ancestors in nobility tracing back to pre-1500, documented in various more or less reliable data bases, IOW they are all more or less cousins of Henri VIII, it's simply that the paths have to be double-checked and added to WT. Once it's done for a few of them, given the strong endogamy of French aristocracy, they will appear at the correct distance of the tree center. Just a question of time, and having more French editors on board :-)
by Bernard Vatant G2G6 Mach 5 (55.6k points)
This sort of thing appeals to the nerd in me. I believe I will see the day that we'll regularly see French people within 20-25 degrees of the connection anchors. There are several noteworthy connections that haven't been built out (and I not being French, I don't feel qualified to do it myself), such as:

Elisabeth (Pelletier de Chambure) de Rothschild, who's parents have profiles but no further:

Lili Damita, who was married to Errol Flynn, who's exploits alone could bring the entire French nation into range of the Tree:

Violette Szabo, the famed British spy of WW2, had a French mother:

Fanny Kreilssamner, who may have had a child fathered by Napoleon, married one of the Duke of Wellington's colonels, and whose daughter married a British aristocrat. Oh, and her sister had a daughter by famed writer Alexandre Dumas. No big deal:

Thx Jessica for those hints. We have a growing number of good connectors between French grandes familles and the Anglo-American Tree, such as the following.

Maybe there is already, or we could create a special category or project page, whatever, a place to gather all those "international connectors".

[edited] ... and I forgot my best connector friend, Richard de Dampierre  who is only 15 degrees from Henry VIII. The Dampierre family is huge, and in the 19th century a few degrees from just about any grande famille in both aristocratie and bourgeoisie.

Thanks for the hint about Fanny Kreilsammer - her sister has a profile, and I'll add linking the two sisters together to my to-do list. I would indeed welcome more hints of possible connections being built out.

The profile for Elisabeth Pelletier de Chambure is unfortunately badly misnamed.

Concerning Violette Szabo her mother's ancestry is already there up to her grandparents. I suspect her chances of connecting would be improved by exploring her father's side.

There are a few more good connectors between France and America. Emily Ridgway is one example. There are more, I just can't find who right now.

Another fun connection: actress Jacqueline Porel (grand-daughter of Réjane, wife of François Périer) had a half-sister France Bellanger, who married George Fitzgeorge, great-great-grandson of King George III.

So I've been looking at Kreilssamner. While her profile says that "she claimed to have been widowed when her husband was killed at Waterloo" and "rumour has it... that her eldest child was fathered by Napoléon I"... I don't know if she was widowed at Waterloo, but her marriage to Lazare Mayer was real enough and took place in Lyon on 22 October 1813. As for her child... We'd have to know the real birth date. (Right now her birth in 1823 rules out Napoléon as father, and obviously also rules out Mayer if he was killed at Waterloo). Napoléon is (like Louis XV) a typical example of "on ne prête qu'aux riches"; he did have a number of children born to mistresses, but certainly not as many as is said. For example, Emilie de Pellapra has been disproved as daughter of Napoléon, who seems to have been nowhere near her mother at the time she was conceived. Similarly, several of the illegitimate children traditionnally attributed to Louis XV don't add up with what is known of the way is private life was organized. 

+2 votes
There has been a lot of work done these past two or three weeks to improve Saint-Ex's connections. Is it better now? I have no ability to make such details statistics as shown on the original post.
by Isabelle Martin G2G6 Pilot (452k points)
It is better now. WikiTree+ sees the two of them now at 32 steps connected. That is less than the 39 he complained about.
If you check Shawn's reply earlier in this thread, you will see he intended no criticism.
I know he didn't, but I can tell you that after all the effort that went into creating that first connection path, it initially felt like a blow.
I understand, Isabelle, and of course anything written in text is going to come across as harsher than it would when spoken. But I "nerd out" over things like that, too, like figuring out how "central" to the Tree someone is, or how many connections to a given anchor, etc. I don't think anyone blames anyone else for the remoteness of the connection. It's not like you can hop in a time machine and go back to convince Antoine de Saint-Exupéry to marry someone else or whatever.

Hi Isabelle, long time no see :) I just ran some tests and it does look like y'all have done great work! Saint-Ex is much more heavily connected into the Main Tree than 2 years ago when I last looked. The average distance from him to all other people on the main tree is now 31.7 (vs. 43.2 in 2018).

I think this diagram makes it most clear: 2020 diagram 2018 diagram. To read these diagrams: Saint-Ex is at the node at bottom, each arrow is the fraction of people on main tree whose shortest path to Saint-Ex is through this link. So 1.0 means basically whole tree goes through these links. If someone is really well connected into the tree in multiple ways, this branches out really wide very quickly. If they are narrowly connected via one line, there will only be one node at each level for a while (like you can see in the 2018 version). In the 2020 version this branches out much more quickly after only 5 steps.

Thanks, Shawn. I guess this sort of schema with a line of several nodes before branching out will typically happen when an isolated profile is created and then "we try to connect it". As opposed to a profile that was naturally created as part of an already connected family.

What is interesting with the diagrams is that there are in fact several paths of connection that don't appear in the second diagram. There's a path (not visible with this week's connection anchors) that goes through Saint-Exupéry-3, Giraud_d'Agay-1 (same as in the diagram) but then Giraud_d'Agay-3, Lallemand_de_Driesen-1, Montgolfier-21, ...  So, what explains that it does not branch out after Giraud_d'Agay-1?

There are other connections (but less good) through Saint-Exupéry-2. (That was the path of the old connection). Just curious, of course.

Isabelle, I didn't mention this before, but I arbitrarily trimmed the tree so that any edge < 0.05 would not appear. So, my guess is that less than 5% of people on the main tree (which is a high bar, almost 1 million profiles now) have their shortest connection through Giraud_d'Agay-3 or Saint-Exupéry-2. This is also using the weekly data dumps so it ignores both: (1) any connections through private/living profiles and (2) any connections made since last week.

I was able to make an improvement recently by connecting Saint-Exupéry's branch of the Taschereau family to the Legardeur de Tilly family.  Using this I was able to decrease my connection pathway to Saint-Exupéry from 29 degrees down to 21 degrees (at least according to the WikiTree+ Connection finder).  I'm not sure of the overall magnitude of the impact, as the featured profiles this week use a combination of the old and new paths.
Even my own connection path goes through that new Taschereau/Le Gardeur connection, so I'd assume the impact was significant. I'm at a hefty 27 degrees though, which - though I notoriously have extremely poor connections - probably means there is still room for improvement. Perhaps lesser French nobility would be worth investigating.

Related questions

+12 votes
5 answers
+17 votes
5 answers
428 views asked Jun 29 in The Tree House by Aleš Trtnik G2G6 Pilot (559k points)
+12 votes
3 answers
+13 votes
4 answers
+25 votes
8 answers
+2 votes
1 answer
+53 votes
22 answers
+1 vote
3 answers
190 views asked Nov 3, 2019 in WikiTree Tech by Ron Johnson G2G6 Mach 3 (30.8k points)
+18 votes
32 answers

WikiTree  ~  About  ~  Help Help  ~  Search Person Search  ~  Surname:

disclaimer - terms - copyright