The scare‐quotes are because “correlatives” is a misnomer. In classical grammars the term traditionally referred to words that are commonly used in pairs, like both X and Y or neither X nor Y (to “correlate” X and Y). Where we'd say “as X as Y”, Latin had a correlative pair tam X quam Y, and it was equivalents to those words – along with a whole bunch of indirectly associated others – that Zamenhof chose to label as korelativoj. An alternative term used by some modern Esperantists is tabelvortoj = “tablewords”, for understandable reasons:
ĉia “every kind” |
ia “some kind” |
kia “what kind” |
nenia “no kind” |
tia “that kind” |
ĉial “for every reason” |
ial “for some reason” |
kial “why” |
nenial “for no reason” |
tial “therefore” |
ĉiam “always” |
iam “some time” |
kiam “when” |
neniam “never” |
tiam “then” |
ĉie “everywhere” |
ie “somewhere” |
kie “where” |
nenie “nowhere” |
tie “there” |
ĉiel “every way” |
iel “somehow” |
kiel “how” |
neniel “no way” |
tiel “thus” |
ĉies “everyone's” |
ies “someone's” |
kies “whose” |
nenies “no‐one's” |
ties “that person's” |
ĉio “everything” |
io “something” |
kio “what” |
nenio “nothing” |
tio “that” |
ĉiom “all” |
iom “some” |
kiom “how much” |
neniom “none” |
tiom “that much” |
ĉiu “everyone” |
iu “someone” |
kiu “who” |
neniu “no‐one” |
tiu “that one” |
Similar schemes occur in some natural languages – Japanese, for instance, has one with four columns and six rows where “that direction” is sochira, “that place, there” is soko, and “that one” is sore – but they occur rarely enough that I've seen it suggested that languages might even be avoiding them for practical reasons: the more similar your question‐words sound, the more likely they are to get muddled. As it is, the generally recommended approach to learning these Esperanto words is to ignore the table and simply memorise each one as an independent vocabulary item. Mind you, knowing Russian helps when it comes to learning the pattern kiel/neniel/tiel = kak/nikak/tak.
These words may form a neat tabular arrangement, but they clash wildly with the standard rules of Esperanto morphology:
- Yes, it looks as if they're constructed by combining prefixes and suffixes, but no, those don't exist as official derivational morphemes – for a start, if they did then the examples mentioned in the Fundamento ought to contain extra “small lines”. The word ia isn't a root I‑ plus the adjective suffix ‑A, ie isn't the corresponding adverb, io isn't the noun, and iu definitely isn't an imperative verb!
- Nonetheless, apart from that last type they all take inflections as if they really were members of the word‐classes they mimic. This even applies to ie, which inflects for case in the crazy way Esperanto adverbs do: ien = “towards somewhere”.
-
The ‐es ending, which looks as if it should be a fifth
verb inflection alongside ‑as/‐is/
‐os/‐us , is in fact a freakish genitive case reserved for the use of particular demonstrative pronouns. “My feet” is miaj piedoj, “one's feet” is oniaj piedoj, and the reflexive “his/her/its/their own feet” is siaj piedoj; but “someone's feet” is ies piedoj, and the only way of saying “something's feet” is to fall back on la piedoj de io! - The words ending in the noun suffix ‑O mystifyingly inflect like nouns for case but not number, so there's no tioj = “those (things)”; and even though the pronouns tio and tiu are distinguished by whether they refer to things or people, tio can't occur as a determiner (before a noun, as in “that book”), so tiu covers both senses (it's tiu libro).
- The words ending in ‑om are used not only as adverbs but also as the basis for noun phrases: iom da libroj = “some books”. These can occur as direct objects but can't be inflected as such; and if we don't need case‐marking to say sep havas neniom da faktoroj “seven has no factors”, why should it be compulsory anywhere else?
- The ĉiu row is another batch of irregular pronouns; but unlike the ones disguised as nouns in the ĉio row, these ones inflect like nouns to give tongue‐twister forms like tiujn = “those (obj.)”. This raises the question: if pronouns can pluralise regularly by adding ‑J, why did we need to learn a separate third‐person‐plural pronoun ili?
English has just a rudimentary “table” for words like where/here/there, whither/hither/thither, whence/hence/thence… but you may now be noticing that while “there” and “whither” have one‐word Esperanto equivalents (tie, kien), “hence” doesn't fit on any row or column of the table. Many languages have a systematic three‐way distinction between proximal “this”, medial “that”, and distal “yonder” (e.g. Japanese ko‑/so‑/a‑), but in Esperanto there's only a ti‑ column; you have to add the particle ĉi to get proximal forms like ĉi tiuj = “these people”, ĉi tie = “here”. (Except that for some reason the proximal form of tiam = “then” gets an irregular coinage of its own: nun = “now”). The obvious approach to fixing this would be to promote the particle to the status of column‐prefix, but no, the ĉie slot is already occupied! And a similar arbitrary one‐off particle is used to stand in for an “any‑” column: iu ajn = “anyone”.
It may be less apparent to speakers of European languages that another thing is missing: a column for relative pronouns (as in WHAT it is) as opposed to interrogative pronouns (as in WHAT is it?). Compare, say, Hindi, where question‐words systematically begin with K just as in Esperanto, but their relative‐clause equivalents have J. Mind you, given that its regular interrogative particle is ĉu, it's surprising that Esperanto forms all its other question‐words with initial KI – especially when that sequence of sounds is otherwise quite rare in Esperanto roots, occurring (for instance) less than a tenth as often as KO. Zamenhof's tendency to avoid KI is the flipside of his ingrained bias against CU; in the Slavic languages, K historically softened to C before I.
Other things that are conspicuous by their absence include for instance rows for “how often?” and “by what means?” or columns for “the same place” and “elsewhere”. Esperantists have been known to improvise workarounds for some of these, but coinages like alie = “elsewhere” are officially forbidden, since there's no orthodox compounding process here that could legally be extended; ali‑e is just the adverb “otherwise”. The only solution is to coin substitute expressions made up via the standard derivational mechanisms: kelk‐foj‑e = “sometimes”, ali‐lok‑e = “elsewhere”, and so on. Esperanto does after all have regular morphemes covering the functions of most of these rows – in some cases redundantly, with both an “affix” and a plain root. But once this option is on the table (so to speak) it becomes obvious that all the so‐called “correlatives” could have been formed out of an open set of freestanding words meaning “this/that/every…” plus another set meaning “place/thing/kind…”, combining as needed to form expressions like “every thing”, “some place”, and so on.