Standard Average European

[Otto Kretschmer]

Where do those features come from?

Do they come from Vulgar Latin and spread to Germanic langugaes from there?

[vegfarandi]

They don't come from a single language, the features are diffused from multiple sources. French and German are the "most average" which makes sense; they're big, historically powerful and in the middle of the continent. There's spikes of certain features centered in other places that don't go across the whole area which area clearly of neither French nor German (country, not language) origin. For example:
- Icelandic, English and Spanish share the phoneme /θ/ in an almost vertical strip from north to south along the Western edge of Europe.
- Norwegian, Swedish, the Baltic languages, Serbo-Croatian and Slovene form an almost contiguous continuum of languages with pitch accents.
- Southern Swedish dialects, Danish, German, French and Portuguese form a contiguous area of throaty R sounds.
- Greek and South Slavic languages form an area of languages that don't have infinitives.

But centered in Germany/France we get a nearly contiguous area of languages that have:
- indefinite AND definite articles preposed to the noun
- inflected relative pronouns
- experiencers in a quirky case (It. mi piace la musica, Sp. me gustas tu)
- no double negation
- particle for comparison (than)
- subject agreement on verbs
- passives with copula (be)
- perfectives with possessive verb (have)
- phonemic voicing within stops AND fricatives
- at least three degrees of height in vowels (I, E, A)
etc. etc.

Some of these stop short at some point but the closer to the center, the higher likelihood of all of the features being shared.

The origin – the origin is complex. Some features go back to PIE, some back to Proto-Romance, Proto-Slavic or Proto-Germanic and then influenced other branches. Most of the features classified as SAE developed somewhere in Central Europe in the past 2-3000 years and spread from there.

[Otto Kretschmer]

SAE area mostly corresponds to former area of the Western Roman Empire or areas adjacent to it so I guessed most features come from Proto Romance/Vulgar Latin

Most Slavic languages do not have articles, do not use compound tenses, use double negation etc. Proto Germanic also did not have articles or periphrastic tenses

[zompist]

I'd just add a couple things to this.

First, the concept veg is talking about, a set of shared features that cannot be explained by inheritance, linguists call a sprachbund. The classic example is the Slavic + Greek sprachbund veg mentions— it actually includes half a dozen other features besides the lack of infinitive.

So, to the extent "SAE" exists, it's a sprachbund centered on French/German.

The name is terrible, by the way. Whorf coined it as a not-very-well-analyzed counter to Native American language. Haspelmath created the more reasonable modern version, but one can hardly read his list without thinking of exceptions— "European" is a pretty bad characterization of the sprachbund's area.

[vegfarandi]

The name is terrible, by the way. Whorf coined it as a not-very-well-analyzed counter to Native American language. Haspelmath created the more reasonable modern version, but one can hardly read his list without thinking of exceptions— "European" is a pretty bad characterization of the sprachbund's area.

Yes, but it's almost always possible to draw a circle or amoeba shape over the area in which the core features exist. That's kind of what makes it fascinating as a concept and probably why Haspelmath bothered refining the theory. There's definitely something there although it's a bit hard to put your finger on it. I think for us conlangers, the takeaway is that geography matters a lot for language evolution. There is the tree-model of linguistic evolution which we all know, and on top of that there's the "radio ping" model, centered in densely populated areas, that crosses language boundaries.

[Ketsuban]

Whorf coined ["Standard Average European"] as a not-very-well-analyzed counter to Native American language. Haspelmath created the more reasonable modern version, but one can hardly read his list without thinking of exceptions— "European" is a pretty bad characterization of the sprachbund's area.

van der Auwera's name, "Charlemagne sprachbund", better indicates its scope - English ends up influenced by it but somewhat peripheral.

I feel like this has been litigated more than enough (to the point where NativLang has provided a nice friendly précis of the whole thing) so a better question might be okay, so there's a sprachbund in western Europe; what can we do with this information? Can conlangers use it to calibrate how "European" our languages feel? Does it help researchers correct for their unconscious biases when observing and documenting other languages? Is it just an academic curiosity like the Balkan or Mesoamerican sprachbünde?