This was an interesting read:

Articles have developed independently in many different language families across the globe. Generally, articles develop over time usually by specialization of certain adjectives or determiners, and their development is often a sign of languages becoming more analytic instead of synthetic, perhaps combined with the loss of inflection as in English, Romance languages, Bulgarian, Macedonian and Torlakian.

Joseph Greenberg in Universals of Human Language describes "the cycle of the definite article": Definite articles (Stage I) evolve from demonstratives, and in turn can become generic articles (Stage II) that may be used in both definite and indefinite contexts, and later merely noun markers (Stage III) that are part of nouns other than proper names and more recent borrowings. Eventually articles may evolve anew from demonstratives.

Wondering if one could elaborate on this. I don't quite understand the second paragraph what some example sentences would look like for each stage. In this context, not sure what a demonstrative would be, vs. a generic article, definite/indefinite context, and noun markers.

I don't quite understand what follows as well:

Definite articles typically arise from demonstratives meaning that. For example, the definite articles in most Romance languages—e.g., el, il, le, la, lo — derive from the Latin demonstratives ille (masculine), illa (feminine) and illud (neuter).

The English definite article the, written þe in Middle English, derives from an Old English demonstrative, which, according to gender, was written se (masculine), seo (feminine) (þe and þeo in the Northumbrian dialect), or þæt (neuter). The neuter form þæt also gave rise to the modern demonstrative that. The ye occasionally seen in pseudo-archaic usage such as "Ye Olde Englishe Tea Shoppe" is actually a form of þe, where the letter thorn (þ) came to be written as a y.

So a definite article is something like the. This derives from a demonstrative like that. The demonstratives in Latin are listed above for example (ille, etc., not sure how many there are, if they are huge in number with conjugations like verbs, etc.). So that gets us [from Stage 0] to Stage 1. But not sure what Stage 2 and 3 are. Wondering if one could explain this in more depth. Hopefully from that it will provide a deeper understanding of the meaning and purpose of words like the and a.

  • The purpose is clear without considering the diachronic process: to mark definiteness or specificity. You've already asked about that before. Better to just ask about the process.
    – curiousdannii
    Sep 18, 2018 at 14:04
  • Not sure what you mean by ask about the process.
    – Lance
    Sep 18, 2018 at 14:07
  • I meant just keep this to asking about the cycle through the three stages.
    – curiousdannii
    Sep 18, 2018 at 14:08
  • Note that par. 3 says that equiv. 'the' typically comes from equiv. 'that', while par. 4 says that English 'that' comes from 'the'.
    – amI
    Sep 19, 2018 at 20:39
  • 2
    @amI se, seo, and þæt are all demonstratives (approximately meaning that). Modern that descends from Old English þæt. Modern the is descended from se, seo, and þæt.
    – CJ Dennis
    Oct 17, 2019 at 4:54

1 Answer 1


Well, I explained the why it's useful in your other question so if you're asking about the process as curiousdannii said, that is you are asking about the grammaticalisation cycle, I could explain a little about the process.

  1. Latin has a rich case system allowing free word order which is used as as a discourse marker; words encoding new or salient information tend to go at the end of the sentence)

  2. For unrelated reasons, the case system collapses, so that word order becomes more rigid, which means there's pressure to find another way to solve the problem which was solved by the flexibility in word order before. Demonstratives are a natural candidate

  3. The meaning of demonstratives weakens because they are now being used much more. So they are generally phonetically unstressed.

  4. We now have unstressed morphemes adjacent to semantically loaded lexemes. These have a strong tendency to cliticise, which is what happened in Romanian and others.

A similar story goes for Germanic languages too, and probably other language families.

  • "the case system collapses" that is so interesting! I would love to know where to learn more.
    – Lance
    Sep 18, 2018 at 16:00
  • @LancePollard Are you asking about Latin in particular, or in general? I can tell you all about the Latin one but I know less about the other languages which have collapsed their cases.
    – Draconis
    Sep 18, 2018 at 18:05
  • @Draconis Any, but yeah Latin sounds good. Latin is a good example one I guess.
    – Lance
    Sep 18, 2018 at 18:10
  • 1
    I thought you would have meant "solved by case endings before", then I'm not sure what you mean here. Perhaps some example sentences could help, for instance to answer the question what kind of things are expressed by flexibility in word order in step 1. There also seems to be the implicit development that when case endings disappear, word order becomes more rigid because it has to provide information about grammatical roles. That, by the way, is not universal. It worked that way in Latin, but in e.g. North-West Semitic a different system appears after case endings were dropped.
    – Keelan
    Sep 19, 2018 at 6:42
  • 1
    @Keelan What Wilson is saying is that because word order was free, grammatically, to be anything the author wanted, they could rearrange the order of words at will to emphasize certain aspects of the sentence—for example, Roman authors tended to put “new or salient information” (or, in my experience, very often the verb) at the end of the sentence. When endings got simpler and declining everything went away, word order had to become more fixed to indicate the things that used to be indicated by endings, but that meant you couldn’t freely reorder the sentence for emphasis. Alternates developed.
    – KRyan
    Oct 17, 2019 at 1:44

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Not the answer you're looking for? Browse other questions tagged or ask your own question.