Among the endless fractal conflicts over the boundaries of concepts (such as whether a hot dog is a sandwich) are innumerable arguments over whether two languages are variations of one language or whether they are distinct, separate languages. From what I have seen these discussions are almost uniformly hopelessly poisoned by the motivating issues of politics and identity. But I’m interested in what it would look like to objectively answer these questions.
The main indicator used to judge whether two languages are the same is called mutual intelligibility. If I can understand what you’re saying, and you can understand what I’m saying, then it’s hard to claim that we are not speaking the same language. (The same concept applies to the written word.) The word “mutual” is there because, for example, Dutch speakers can understand Afrikaans speakers better than Afrikaans speakers can understand Dutch speakers.
But immediately following from this invitingly crisp designation is a degree. How much, exactly, did you understand? For this, I think it’s reasonable to take some standard corpus within the language, like the nightly news broadcast. After all, there are plenty of exclusively English speakers such as James Joyce that I cannot understand, due to, um, their being bad at making sense.
In addition to the blurriness of how much one understands, here’s another dimension that I find more interesting: how long it takes to understand someone. Say that you meet someone and you can immediately tell that they are “from somewhere else” (namely somewhere different from you). They have a very “thick” accent. The first few words out of their mouth you will probably not understand. So in that case the intelligibility is zero. But then as they say a couple more sentences, your brain somehow catches whatever vowel or other phonemic shifts they’re using, and you’ll start to understand. There will still be some problems understanding them for a while, as each new word they say has a chance of mismatching in your brain. They may even have a lot of regional vocabulary that you couldn’t figure out without being told. But it will pretty quickly converge to something that you would mark as a “yes” under the metric of mutual intelligibility.
When I visited Paris, I found that I could read a surprising number of signs. I couldn’t understand a single thing anyone said, but the signs were helpful. It turns out that French and English have a fair number of cognates, and so when you pass the sign for “Musée de l’air et de l’espace,” you can know what you’re about to experience. I also have a fairly large technical vocabulary, much of which comes from Latin, and since French descended from Latin, I had higher recognition.
So French has at least some measurable degree of intelligibility to English speakers. And most of the cognates (recognizably related vocabulary) don’t come immediately. For example, the word for “left” is “gauche”. Yes, like the word gauche as in awkward. Being left-handed is “wrong”, socially non-compliant, awkward. This connection does not come after just a few more seconds, as in the opening example of listening to a person’s accent, but it comes relatively quickly, and you can often remember the meaning after only one exposure to the cognate connection.
So intelligibility can be measured in something akin to a half-life: how long it takes before you can understand X% of what they are saying. This time duration is really the length of time you tried to understand, either by seeking patterns, being told patterns, or otherwise deliberately learning things about the language.
On the far side of dissimilarity, Japanese is profoundly and fundamentally disconnected from English in the great tree of languages. It may be that spoken language first occurred at one place and time, and that all languages are ultimately descended from it, but that is not known, and for the purposes of learning and communication, an English speaker can only learn the meaning of a Japanese word by asking, and never through cognates. But even then, learning time is a good measure of similarity. The more novel grammatical structures you have to learn, the more different of a language it is.
I think Scots is a good illustrative case for this measure because many speakers of English cannot understand literally anything a Scots speaker is saying right away. It is much harder to understand than for example a strong US Southern accent. But if you pay attention and try to connect things up to context et cetera, you can, on the order of tens of seconds or a few minutes, catch the phonemic shifts and then understand more like 50% of what is being said. You can then learn some of the more common word replacements, like how “ken” means “know” (the shared k and n are not a coincidence) or how “twa” is “two”. Picking up on the negative contractions is especially helpful: “dinna” for “don’t”, “havenae” for “haven’t”, or “wadna” for “wouldn’t”. Altogether I suspect the differences put its intelligibility “half-life” at a few hours of study to be able to understand the equivalent of a nightly news broadcast. This is very different from a thick accent, and also very different from needing to take a year of classes to understand.
It is the perfect degree of distance to incite maximal argumentation over whether it is the same language or not.