search

CreoPhon / consonant_stability_portuguese_creoles

Supplementary materials for Silva & Moran, "Consonant stability in Portuguese-based creoles", to appear, Diachronica.
0 stars 0 forks source link

Stability and duration of contact #3

Open silva-carlos opened 2 years ago

silva-carlos commented 2 years ago

Code and plot the correlation between the duration of Portuguese influence and the stability values.

bambooforest commented 2 years ago

@silva-carlos -- is this "End of influence" - "First major settlement"?

Some of these go until 1999 or so. Others not so long. Are these fair comparisons?

bambooforest commented 2 years ago

I've added the plots for duration vs overall creole stability:

https://github.com/silva-carlos/consonant_stability_portuguese_creoles

Does not look like there's anything there. Should we look into individual segments or segment classes?

silva-carlos commented 2 years ago

@bambooforest , you're right: "duration" = "end of influence" - "first major settlement". The values for "end of influence" correspond either to the year of independence or to the year when another European nation took control over the colony.

bambooforest commented 2 years ago

@silva-carlos -- let's make a document that describes the variables like this, e.g.:

silva-carlos commented 2 years ago

So, we found that there is no strong correlation between overall stability and duration of contact. That's an important discovery! We can look into individual segments or segment classes, of course. I will think about some hypotheses that we can test. It would be interesting, for instance, to evaluate the extent to which creoles with a longer contact follow (or do not follow) the evolution of their lexifier.

bambooforest commented 2 years ago

A potential problem here is the duration of contact situations recorded varies greatly. There is also a bunch of NA in the mean stability values.

I've added a table here to the README file. Please have a look.

Do you know why offhand there are NAs?

You also had a recent course on linear regression, if I recall correctly. Perhaps you could add a few models here and see if there's anything beyond the eyeballing that I've done?

For example, maybe it make sense to test whether stability versus duration matters in languages that have similar time depths of occupation?

All in all though, there are too few data points probably. Or?

silva-carlos commented 2 years ago

Thank you for pointing this out, @bambooforest!

I can see that there are two reasons why this is happening:

(i) the values of "PlaceStability" and "MannerStabiltity" (0 or 1) have not been assigned yet (Patua Macau and all Spanish creoles); (ii) we have located a creole but didn't collect the data yet (Timor creole).

We really need to decide what to do concerning the Spanish creoles (e.g. exclude them or put them in a different file, etc).

That would leave us only with Patua Macau and Timor creole to deal with. I will add the missing data so we can make this work and move forward.