Exploring Galapalogos

Exploring Galapalogos

or is Olivia on her way?

Jezebel recently released an article on the most popular names for girls over the years as gathered by the us census
(sidenote: i’m so happy that this data exists).
Article

Lots of people have discussed the aspect of conformity and or how common one’s own name is around this, but I think that not nearly so fascinating as the structure and dynamics of the transmission/infection of names from state to state.

Unfortunately Because of the way that they colored the diagram this doesn’t jump out at you unless you are primed to look for it. Apparently by studying the evolution of ideas etc. I’m in the camp of “primed to look for this”.

The rest of this will proceed with some non-mathematical descriptions of some evolutionary theoretic ideas.
(These ideas are well discussed in a similar non-mathematical way in the following wikipedia entry about the Galapagos Syndrome in Japanese 3G cell phones).

And then I will attempt to apply the theory to the naming data, though admittedly only cursorily.

An high-level evolutionary theoretic primer on Galapagos Syndrome

  1. Semi-isolated populations will be able to have mutations occur in them and prosper in them that are not well represented in the total population of a species.
  2. Larger populations that are well integrated will have fewer shifts.
  3. If there is variation in the population you expect that variation to peek through more readily in smaller, less connected populations because it takes that much more to oust the ruler of a larger kingdom (even if more representatives of dissident factions exist there).
    1. It is useful to note that a new species and a variation on an existing species is not always an easy line to draw. However evolutionary theory tends to speak in terms of hard categories of species, i.e., the propagation of category A is equivalent to the species “A” growing in a population, despite the fact we often observe a good deal of variance between instantiations(individual organisms) of species A. (e.g., see linklist if you want 6 examples of 1 example of individual differences between instantiations of a type)
  4. You would not expect the deviations from the type to be able to survive if they are exquisitely adapted to a single environment so much so that they became their own type.

Calling the Namegame

Now look at the images from that slide show.
He’s a brief description using my brain as a statistical filter (its biased toward large areas with different colors), with the rise of a horse race style description as I get bored of merely listing facts.

  • 1960s: Fall of Mary, Rise of Lisa, Fail of Brenda, Floundering of Michele
  • 1970: obliteration of Lisa. tries to survive as melissa. Angela tries and fails
  • 1970s: Jennifer takes over the entire country for the entire decade almost literally…until 1983 Jennifer wins. Deviations described below
  • 1979: Amanda tries really hard in the South, and dances around for the next decade or so. kinda sneaky, i appreciate its effort e.g., New Mexico(NM) and North Dakota(ND) 1988, New Hampshire(NH) 1989
  • 1980: first sighting of Jessica in Alaska(AK), Maine(ME) and NH
  • 1981: Jessica still hangs around in ND, ME, NH, Rhode Island(RI). AK back to Jennifer
  • 1982: Jessica: Montana(MT), ND, Kentucky(KE), West Virginia(WV), NH, ME, RI
  • 1983: New sighting — “Ashley”! Louisiana, South Carolina, Tennessee, Indiana; Jessica: ND, Wyoming(WY), WV, NH, ME
  • 1984: Jennifer loses its stranglehold. Duopoly of Jessica/“Ashley” arise
  • Jessica: more remotish states (e.g., AK,MT,NM,CO,SD,Georgia(GA),WV,VA,ME)
  • 1985: Jennifer topples, Illinois Massachusettes and Deleware are holdouts
    Jessica has arisen!
    “Ashley” has arisen! (they really are equivalent at this point)
    Jessica has gateway states.
    “Ashley” has heartland states + Hawaii(HI)
  • 1986–1989 Jessica, Ashley Duopoly
  • 1989: Rise in Brittany in the south (side note: I wonder how the census dealt with spelling variation[1])
  • 1990: Brittany lingers on, still Jessica Ashley duopoly though.
  • 1991: Ashley takes over, Jessica still around, Brittany limps along, and Sarah emerges(NH)
  • 1992: Jessica and Ashley again, Brittany all but dead (WV), Sarah(NH), Amanda reemerges (RI) so sneaky!
  • 1993: Jessica technically dominates, but Ashley has apparent geographic plurality, Sarah relocates (MA), Samantha seen in ND, Emily arises in the northeast (ME,NH,Vermont,Connecticut)
  • 1994: Jessica pushes back, takes over Ashley territory, Samantha moves to Wisconsin(WI), Taylor appears in ND and LA, Megan in SD, Brittany raises her head in WV, Emily spreads to the midwest (Minnesota(MN), and Iowa) in addition to its NE origins + Pennsylvania (PE)
  • 1995: Emily taking over more ground, Jessica still reigning, Ashley fading, Samantha appears in AK and RI, Brittany holds strong in WV, Taylor moves into more of the south (GA and South Carolina (SC))
  • 1996: Pluralism makes it hard to describe! Emily winning, Ashley finally in AK…Hannah arises (south), Taylor lingers(SC), Brittany fades away Emily in WV, Alexis in LA, Madison in Kansas(KA) and Oklahoma(OK), Samantha in NM SD and New Jersey(NJ)
  • 1997: Hannah takes over south, but for Alexis in LA and Ashley in Florida (FL) Hannah also in AK, MT, CO, SD; Taylor HI, NM & OK; Samantha New York (NY), Madison in Utah (UT)
  • 1998: Alexis spreads to Mississippi(MS) and Arizona(AR) from LA, Samantha to NY and NJ, Ashley’s last stand in NM; Madison in Idaho, WY, UT and KN, HI still holds out on Taylor. Emily still winning
  • 1999: Hannah makes a play for the bigtimes: Ohio (OH), CO, ND, IN, KY, WV, TE, North Carolina (NC), SC, Arkansas (AS),Alabama(AL), LA, NH; Madison in MS, MT, KA,OK; Samantha Nevada(NV) and AR; Taylor still in HI and now WY
  • 2000: (note the graph is miscolored on TE) Its a Hannah Madison Emily free for all. HI shows some Galapagos effect of uniqueness with “Kayla”
  • 2001: Emily still technically winning, but Madison with the geographic plurality (including AK), Ashley resurges in AR Alexis holding strong in NM. HI goes with Taylor
  • 2002: Madison gains more ground, Hannah relegated to SD, Emma appears in MI, WI, NH, Nebraska (NE). Kayla in HI again. Alexis still in NM.
  • 2003: Emma has taken over the North + Hawaii. Madison left to the South East and Emily to the larger states and South West and some of the North East. Hannah jumps to AK, Alexis lives on in NM.
  • 2004: Madison Emma Emily free-for-all. Emma has both HI and AK. Ashley reappears in NE. Alexis shifts to Alyssa in NM. Emily technically wins (populous states)
  • 2005: Madison loses ground fortified around the East coast of Ohio river and the Mississippi river once they merge + CO DE and AK. Alyssa holds strong. Ava seen in MN. Olivia sighting in CT & RI, rest are Emma and Emily. Emily wins (populous states)
  • 2006: Isabella appears NM, CO, HI, FL, NJ, CT, RI; Ava sweeps across the great lakes MN,WI, MI, PA, + ND & MA. Mia pops up in AR. Emma and Emily do battle, Madison gets a strong foothold in the south East up to Ohio River. Emily Still wins (populous states)
  • 2007: This year is a hot mess. First Emily wins; carries 3 states California (CA), Texas (TX), NV. Olivia shows up again in UT VT and Washington (WA) Isabella includes AK and AR, IL, NY, + those from before(I’m honestly surprised it didn’t win this year). Emma makes an incursion into the Madison stronghold TE & AL. Ava spreads to OH, IO, SD, NH, DE plus where it was before, lost ND to Emma. Madison jumps to MT & ID, and in the process loses a phoneme gains in math with Addison taking over NE and KA. Sophia appears in HI, RI. I remind you Emily won with 3 states.
  • 2008: Olivia makes a stand in the northwest. picking up OR, ID, UT to go with WA as well as MI and IL; Emily still in NV and TX, loses main title. CA goes to Isabella. Emma continues to migrate into the Madison territory, takes the title. Madison still exists in patches. Ava Still going strong on the great lakes + MA + NE + DE; HI ever the wildcard goes with Chloe. Emma wins with a plurality
  • 2009: Isabella sweeps! new: TX, OK, WY, NV, WA, AK, IL, OH, WV, VA, PA, DE, HI. Olivia takes over most other Ava states (new: ND, MN,WI, MI, VT, MA). Addison appears again(NE). Ava remains in IO spreads to LA. Rest split north to Emma, south to Madison. 2009 goes to Isabella.
  • 2010: Sophia reappears in North, AK, WA OR, ND, SD, WI, MI NH, VT, Maryland (MD), DE. Ava adds MN back to IO and LA. Olivia curtailed to ID, UT, NE,VA. Emma holds out in MT, AL, Arkansas(AS),NC,Indiana(IN), ME. Madison holds onto MS & SC. Rest goes to Isabella. 2010 goes to Isabella
  • 2011: Sophia and Emma carve up Isabella! Isabella holds onto FL and WV. Madison holds onto MS and SC. Ava in SD & LA. Sophia gains West and most coastal areas (all of West coast, WI IL for great lakes, Northeast area from MD up minus ME and NH), Olivia has AK, UT, CO, MN,MI. Emma has the rest. Sophia wins!
  • 2012: Emma topples much of the competition, but Sophia remains triumphant holds onto West Coast (CA, WA,OR) + larger states(TX,IL,OH,NY), Some of the South west (AR, NM), some of the East Coast RI, NJ, MD, DE, VA. Isabella grasps onto FL. Ava stirs in NH. Olivia basks in ID. Everything else is Emma’s Sophia wins.

Gaming the game with evolutionary theory

There are some interesting trends that deserve investigating.

Nation/time wide:

  1. We are more fractioned, name wise. We’ve seen nothing like the ubiquity of Jennifer in the 70s since, well Jennifer in the 70s.
  2. Some names hold out for a long time staying in the same state for a long time (e.g., Madison) or changing states frequently and appearing again and again even after breaks of not being present (e.g., Amanda, Hannah, Ava, Ashley)
  3. In some cases names arrive with little warning (Emma, Sophia), or with completely no warning (Jennifer).
  4. Large population states sway the “winner” a lot. See 2007, Emily.

State wise

Hawaii is weird: lots of unique names app are and go no where. New Mexico is similarly weird. (see evo point 4). However they do differ.
However, Hawaii compared to New Mexico is a much better indicator as to where a name will go. This may be because it is more isolated and the fact that New Mexico has such names while being close to other states suggests that the uniqueness of the name climate helped spawn and constrain both Alexis and Alyssa. On the other hand Hawaii had Isabella and Sophia 3 years early. But also held onto Taylor for a while while no other movement was happening. Kayla and Chloe are yet to be observed elsewhere.

Alaska on the other hand (with Maine and New Hampshire) successfully housed Jessica as Jennifer’s reign continued, though Maine was the most staunch of those supporters.
Maine though has a tendency to hold out for a while and then jump ship shortly after the name Maine held became too popular.

And in truth even as I try to make these generalizations, none of them seem to hold water beyond merely being post hoc analyses. Except for the general trends. Smaller states seem to undergo more variation(see evo point 1). Large states(CA, NY,TX,FL) with little isolation (most are large populous states with many ports), change slowly and rarely(see evo point 2). Look to smaller states to observe the endemic names that are just waiting to spring up(see evo point 3).

This is in line with some of the evolutionary theory regarding small isolated populations and large integrated populations. However, as I had originally put it small and isolated were treated as equivalent, but obviously that need not be the case. The question is where one draws the line around a population that is small but well integrated — why is it not part of the larger population together.

Soo…..what?

Of course all of this is just my qualitative, eye-balling of these data, to be able to do this analysis justice I’d need at least the following things (aside from more quantitative data more generally).

A point on names and spelling

There are many ways of spelling Brittany [1], and many of these names. I’m curious how the census decides to collapse alternate spellings. Those would be more akin to mutations of the sort that could give rise to the Madison → Addison shift.(see evo point 3.1) Might some names persist better because there are fewer variations on the spelling? Or would that make them persist worse? What does it mean for a name to persist if it only persists as a wide set of variants?

A point on distributions and point-estimators

So far I only analyzed the yearly point estimator. However, knowing what we do know about how names work, my guess is that in many cases (surely not all) we could have done a better job predicting the rise and spread of names if we knew not just the top name, but rather the actual distribution of names, or even just the top 10 names. I’m sure that graph would be horribly unwieldy though.

A concluding whimsical prediction

I could very well be wrong about this, but…

For all those expecting (or soon to be expecting) parents out there…
Is it Ava or Olivia who is on her way?

  1. A regularly express Br(i|y)(t|tt)(e|a|i|y)?(n|nn)(e|a)?(y|i). That may be too permissive (e.g., it allows Brytynnai) but it gets the idea across.  ↩