As I learned along this path,

Looking at "languages" as they are spoken today, some observations are obvious enough to be trivial: they vary, they change, and they are abstract concepts that refer to some largely arbitrary collection of varieties that vary and change simultaneously.

What we call a language is essentially made up of mutually intelligible idiolects and social conventions and philosophies. Even mutual intelligibility is socially defined at every level of varieties and various "-lects" we talk about. For example, a speaker of a Northern American English dialect is likely to fail to understand Kenyan or Indian Englishes, while speakers of Portuguese and Spanish and Italian are likely to understand each other alright, when non-border varieties of these languages are much more significantly different from each other than the mentioned Englishes. The difference is easily explained when you consider that speakers of the second group of languages are often exposed to the use of each others languages, when a Northern American English speaker will rarely interact with a speaker of Kenyan or Indian Englishes. The why of this difference is beyond the scope of this post but suffice it to say that it has to do with power relations and histories of subjugation, and thus social matters.

So, the term "language" is essentially intractable in a coherent manner when referring to the actual linguistic material they use to communicate with each other. It is a political term, fit for use in that context, but in linguistics we can only accurately use it when we speak about that very politics, "language ideologies". Of course in this light, and "dialect" and "accent" existing in very similar situations, in linguistics we've taken to using terms like "variety", "code", "speech", "sociolect", and some other terms with the "-lect" root.

On the ground, speakers command a whole array of "varieties", which they use, sometimes discretely, sometimes mixing and matching-ly. In the study of language, we often consider individuals to have at least one variety, and we consider these situational variation as rule based divergences, or as multilingualism, in which each variety a speaker has also has these situational variants, e.g. formal and informal registers, written versus spoken variants, and so on.

Maybe a better way to put this is that speakers command multiple, changing, interacting idiolects, each affected by a (slightly or largely) different social context.

Let's consider me. I command three "languages" "well": Turkish, my first language, English, a lifelong second language, and Italian, my third language which I learned at university as a student of Italianistics. In each of these languages, tho with different competence, I am able to communicate formally and informally, and I have some knowledge of some "non-standard varieties", or "dialects", and such.

This means that, according to the above model, I have for example a set of idiolects that are influenced by what's conventionally considered to be Turkish. I have a formal idiolect that is influenced by prestige varieties and written language, I have an informal idiolect that is influenced by standard language, by common varieties that I am exposed to thru localised interactions as well as mediatised ones, by multiple other varieties be it dialectal or sociolectal, and then I have a "sincere" informal idiolect that is influenced, along with all of the above, by varieties classified under Turkish that are spoken at home, including urban Istanbulite Turkish that my parents speak, Urfa/Birecik varieties that my maternal grandparents spoke which also affected my mother's idiolects of course, and a similar situation with my father and his relatives who are from the Hınıs area of Erzurum province. Just within Turkish, I am, like everyone else who can speak at least one "language", able to code switch between all these idiolects, often in a self-aware and semantically-contentful fashion. With many other people like me, wilfully or not, I participate in the collaborative, political co-creation of the "Turkish language" as a complex social construct that encompasses varieties and ideologies around them, and all the speakers' social realities as well as judgements about who is and who is not a speaker.

Through these building blocks, and sociopolitical interactions, with the effect of physical and digital/mediatised geographies, languages and dialects as sociopolitical constructs are born.

These varieties, based on their sociopolitical interactions, coexist, influence each other, or exist in antagonistic and/or assimilationistic conflicts, and as outward expressions of identity. They change, merge, split, whether unconsciously or by political action.

Here, because all these ad hoc and vague terms are way too confusing, and because there's not enough overlapping-but-not-quite-the-same jargon in language studies, I will create a term to refer to "any form of language that can be defined in some coherent, useful manner", and it will be "code". For example, "Turkish" is a code because I can define it as "a standardised language maintained by the government of the Republic of Turkey". Further, because it is an instance of a standardised national language, maintained by a government, I'll call it a "rectiocode", from Latin "rectio, rectionis" for government. "Turkish" is also a code at a separate plane as the language associated mainly with the Turkish ethnicity, and in this context what it entails is different from what the rectiocodic use of Turkish does, so in this context I will say I am talking about an ethnocode that can be referred to as Turkish.

"Birecik Turkish" is a code that mainly pertains to the Birecik town found between the Antep and Urfa provinces of Turkey, which means I can refer to it as a "regiocode". The Birecik Turkish regiocode is connected to Turkish the rectiocode and to Turkish the ethnocode in complex ways, as well as to ethnocodes Kurdish and Arabic, and their relevant regiocodes. Further codes like Quranic Arabic, Armenian, Syriac are relevant as well. Every "code" is a nexus: it is influenced by other codes at all levels, from standardised national languages to small neighbouring regional varieties, and in turn influences them, based on the complex societal entanglements of their speakers.

In this fashion we can talk of religiocodes, raciocodes, sociocodes, idiocodes, and even pseudocodes which would be those codes that only exist in abstract fashion. For example, Proto-Indo-European would be a pseudocode because it does not refer to a code that was spoken by anybody, but a reconstructed, model code that represents a pool of codes that have diachronically changed into another pool of codes.

And therein lies the major problem with genealogical/historical linguistics: as a relic of a time when linguistics was chiefly dedicated to much more sinister undertakings, namely the construction and practice of nationalism, colonialism, and racism (which we've nowhere near cleaned our hands from, e.g. SIL, e.g. white supremacist structures and research practices, e.g. Anglocentrism, e.g., e.g., e.g., e.g.), based on an ideologically charged methodology, namely comparative method, the model of language development that we have developed is not only inadequate at representing the above complexity, but actively hinders our comprehension and explanation thereof. Armed with my fancy new terms, let me try to explain this.

We said that codes coexist, covary, codevelop, in timespace. They influence each other, and who influences and to what degree and whom is also variable. And yet our historical linguistic model suggests a tree model, a phylogenetic model, that privileges either a constructed anachronistic code or an imposingly influential historical code. This means that the entropy associated with the life of a group of codes—dare I call it, a codeme—is obscured, and we're confined to thinking and/or expressing that variation lessens as we travel backwards in time.

This conception is inherently tied to the approach to language that considers "language" and "dialect" as concepts that are tractable on a non-political plane—as realities definable, confinable, describable, without reference to language ideologies. About any linguist worth their salt would know to say this is impossible. They'd know to say that a "language is a dialect with a navy and an army". If we consider that, and if we consider the complexities of inter-code interaction, we couldn't but say that this conception is flawed, and that a better model would be primarily aereal and interactional. We could still talk about genealogical connection between codes, perhaps talking about geneacode(me)s, phylocode(me)s. But instead of fronting this diachronical, phylogenetic, and often abstract relation as the main model through which we explain distribution and variation of codes, we would have created a model that's much more adept at expressing the entire gamut of ways in which codes coexist and (co)develop throughout timespace (always with an eye for abstract spaces, such as literary languages, digital spaces, lingua francas, and so on; language is never solely tied to physical space, even in history).

I believe this distinction matters because the current, chiefly genealogical model not only makes it hard to communicate or conceptualise that diachronical pseudocodes/geneacodes are abstract groupings that represent a large collection of codes that are grouped based on a whole variety of factors that range from sociopolitical to ideological to scientific, but it also promotes an understanding of ethnolinguistic history that is dangerously simplistic, suggesting a tree-like history, when the reality is much more complex, much more richer, and much less viable as a building block for injustice and oppression. Genealogical linguistics and comparative method are inextricably entangled with nationalism and nation-building, and relevant ideologies; they are not neutral, scientific methodologies that produce ideologically neutral results.

In that light, I believe that, for ideological and scholarly reasons, it is about time that we advance on from this limited model with a tragic history, and seek more up-to-par and open minded, egalitarian approaches. I have at times seen linguists work in this direction, but my scholarly interests have changed in ways that shifts this concern outside my focus, so instead of taking up this project academically, I approach it here on my blog, philosophically.