Telicity is a grammatical contrast that indicates whether the intended result has been achieved. This adds the so-called telic entity to the sentence - the telic entity has a very strict meaning in the sense that the action must be complete as intended. Without the telic entity, the meaning is vague - it may or may not be complete. One system of denoting this is with the case of the object, as in Finnish: the object is marked as "finished" (telic action), or "unfinished/unknown" (atelic action). Another marking is with a verbal prefix, as in German and Czech, where the action is marked with a prefix, if it is telic and perfective. Estonian uses both systems. Also, the Piraha (Pirahã) language is reported to contrast telicity.
This text is mostly about Fennic telicity; any Piraha speakers are welcome to protest.
Telicity is not the same as the perfective verb aspect (in e.g. Germanic languages or Slavic languages) even though these two often coincide. This can be illustrated by superposition: distinct telic-perfective, telic-imperfective, atelic-perfective and atelic-imperfective clauses can be formatted. It's not grammaticized at all in other languages such as English, and consequently telicity is one of the hardest things to learn properly in Fennic languages.
The term "telicity" is also used in analysis of the English language, but unlike Finnish, there is no grammatical device to mark it, so some verbs are considered always telic (such as to write) and the other verbs are atelic (such as to dream). The telicity of English so-called "telic" verbs is actually contrasted in Finnish; the latter has a direct correspondence, partitive verbs, in Finnish and Estonian — more about them below.
Telicity is not marked grammatically in any way in English, so we have to build the right context to express it. English permits indicating perfectivity with the presence of articles (a/the) in some special cases involving mass nouns. This is not telicity, but can be used to express it using suitable assumptions. Let's assume this: "He tried to drink coffee, and if he succeeds, none is left." After that, we might say:
1. He drank a/the coffee. = He successfully drank coffee. (telic entity)
2. He drank coffee. = He unsuccessfully drank coffee. (atelic)
In example 1, the object is "finished" and there is a telic entity, marked as "coffee-telic". In example 2, the object is "unfinished/unknown", so it's marked "coffee-atelic". In Finnish and Estonian, this is coded with the cases accusative and partitive, in this way:
1. Hän joi kahvin. "He drank coffee, succeeding drinking it all." (telic - accusative) see small print 1
2. Hän joi kahvia. "He drank coffee, unknown if succeeded." (atelic - partitive)
Notice that the accusative ending is -n
, which Finnish sometimes elides, and with the exception of some Finnish pronouns, where it is -t
, e.g. hänet
"him/her", or sinut
"you (singular)". The ending -n
is also a genitive
, so this may cause some confusion. (The Finnish accusative directly translates to a Latin or other Indo-European accusative — but beware, so does the partitive.)
In Finnish and Estonian, every object has to be marked in this fashion. It will distinguish important meanings, as for every action, the listener will assume the intended consequence. For example, here's an enormous juridical difference:
1. Poliisi ampui rikollista. "The policeman shot (at) the criminal." (atelic - partitive)
2. Poliisi ampui rikollisen. "The policeman shot the criminal dead." (telic - accusative)
The object - in this case the criminal
- is finished, and with the verb "to shoot", it means he is dead
. This telic grammar is an integral part of Fennic languages and has to be used every time you utter a sentence with an object.
The same action may not refer to a telic object again. An apparent analogy in programming would be calling the Java method
Timer.cancel() — although the timer still exists, no more operations may be scheduled into it.
The same action may refer to an atelic object again. However, in Finnish grammar, this does not give any guarantees that the object is unfinished. For example, in the previous example, an atelic object for the action of shooting may die. If we indicate this death in a further sentence, this sentence structure indicates that the death is disconnected from the shooting:
1. Poliisi ampui rikollista ja rikollinen kuoli. "The policeman shot at the criminal, after which the criminal died."
Approximately, the accusative translates to "the" or "entire", and the partitive to "some", e.g. "I drank some coffee" becomes "I drank coffee-partitive", and "I drank the entire coffee" becomes "I drank coffee-accusative". This is a rather important realization, because noun and its case is independent of the action. For example, you are asked: "What did you buy", or "what was in there", or anything of that vein. Now, you can answer "some fish", which is partitive, or "a fish", which is accusative. (E.g. Finnish kalan, kalaa, respectively.) Here, the action might be anything, or even nonexistent as with "what was in there". The noun needn't to be an object for any verb at all!
This actually explains the origins of Fennic atelicity. The case marked with -ta is an ablative marker (such as English "from") in Proto-Uralic, and has developed into a partitive in Fennic languages. This is illustrated by the fact that all cases meaning "from" — elative -sta, ablative -lta, excessive -nta — end with -ta. The partitive means "a part of larger whole", where this meaning originated as "from a larger whole". For example, Juon vettä is "I drink (some) water" in Finnish, but its old, Proto-Uralic meaning was "I drink (from a mass of) water".
Interestingly, there is a misconception that the native word for "Finnish language", suomi, could be alternatively written as suomea. This is because if you ask "Do you speak Finnish", unless you're really a babbler, you can speak the Finnish language only partially. This means that the question translates with the word suomi "Finnish language" marked for atelicity with the partitive -ta; the final vowel elides and gives way to an epenthetic vowel 'e'; the 't' of the ending elides as it's intervocalic; thus, we agglutinate suomi+ta → suomea, and get the sentence Puhutko suomea "Do you speak (some) Finnish?" (In the accusative, the sentence would really mean that you'd speak every word and every sentence in the language.)
The telic grammar discussed is an exclusively Fennic innovation with respect to other Finno-Ugrian languages; Hungarian, for example, has the definite vs. indefinite contrast. Incidentally, Finnish speakers sometimes use the word se "it" much like it was a definite article, but this does not interfere with the telicity. Consider this superposition:
Poliisi ampui rikollisen / rikollista. "The policeman shot at (a/the/some) criminal / (a/the/some) criminal dead"
Poliisi ampui sen rikollisen / sitä rikollista. "The policeman shot at the criminal / the criminal dead"
Neither Finnish nor Estonian use a future tense, because telicity can replace it to a large extent. For example:
Luen kirjan. "I read book-finished." - can mean only "I will read a book."
Luen kirjaa. "I read book-unfinished." - can mean only "I'm reading a book."
You cannot read a book immediately. Thus, any references to a finished book means that the action takes place in the future. Likewise, an unfinished book is being read right now. Just as indicating the future tense is mandatory in English, indicating telicity is mandatory in Finnish and Estonian. (Telicity is not the only way to express a "future tense": the potential mood
, a mood not found in English, the conditional mood
, words like "tomorrow", or just by context. Nevertheless, there is no single grammaticized future.)
Some verbs have a scope of meaning that is graduated finer only by telicity. The verb "silittää" (to smooth/stroke/iron) is one. If you stroke something telic (finish it), the action is ironing. Whereas, when the object is atelic, the action can be any of smoothing/stroking/ironing. For example, "Silitän kissaa" (atelic) means "I stroke a cat". Then "Silitän kissan" (telic) gives the hilarious idea of you holding a cat on a ironing board and ironing her.
With some verbs, the Aktionsart (i.e. what actually happens and where) changes its focus on a different telicity. For example, "odottaa" (to wait). Wait something unfinished (atelic), you "wait for something". Wait something successfully (telic), you "wait through something". "Odotan tunnin" means "I wait an hour, successfully", that is, through the hour. "Odotan Niinaa" means "I wait Niina", that is, "I wait for Niina."
Should you mix these up, the context will change, and thus the meaning is torn asunder like a squirrel in a lawnmower. For the first, "Odotan tuntia" means "I wait for the hour", where in this context "hour" means an hour of teaching, landing on the meaning "I wait for the class to begin". The second sentence, "Odotan Niinan", is nonsensical. (How do you completely wait someone?) As the "-n" is also a genitive, the reader assumes this to be a sentence fragment, "I wait for Nina's". (In fact, the ''ära'' construction explained below can be used in Estonian in this context, e.g. ''ma ootan su ära'' "I will wait for you" (telic), i.e. until successful. Thanks eallik!)
There are also verbs for which the meaning is completely, not just slightly, different, on a different telicity, e.g. "naida"; "marry" when telic, "have sex with" when atelic or telic with an explicitly given result that differentiates the meaning by context.
There are verbs that govern only either case. There are so-called partitive verbs, which can only be used in the partitive, i.e. they're always atelic. This is basically a grammatical agreement, because some differences exist between Estonian and Finnish. For example, let's compare some feeling-related Estonian partitive verbs with the corresponding Finnish verbs. The corresponding Finnish verb is marked with an asterisk (*) if it's not an exclusively partitive verb:
armastama "love" rakastaa*
imetlama "admire" imarrella
kartma "fear" pelätä
kiitma "praise" kehua
tundma "feel, know" tuntea*
The one "tuntea" is usually the exact opposite, or an accusative verb in Finnish. (Cf. Finnish tunnustella "palpate", which is a partitive verb.) Some verbs are close to being partitive verbs, particularly such that the context where a telic sentence would be meaningful are hard to imagine. For example, "rakastaa" (to love). You almost always love someone without end, without a defined endpoint. You can, nevertheless, construct a counterexample with an endpoint: "I want to love you to death". This means we mark the word "you" like this: "I want to love you·accusative to death", i.e. Tahdon rakastaa sinut kuoliaaksi (telic).
Estonian has an additional telic system. Estonian uses a telicity-indicating preverb, ära, which basically means "away" or "up", as in "drink up". For example, jooma "to drink", ärajooma / jooma .. ära "to drink up". As with English and Finnish constructions, this can give a different meaning for the verb, e.g. ostma "to buy", but äraostma / ostma .. ära "to bribe someone", like "to buy someone away".
Some Indo-European languages also add a telic entity to the sentence along with the addition of some derivational suffix indicating perfectivity (="action is finished"). Such are Czech pre-, meaning "again", as in pre-psal "rewrite", or High German er-. User izubachi gives this example:
Der Polizist hat den Verbrecher geschossen. "The policeman shot (at) the criminal."
Der Polizist hat den Verbrecher erschossen. "The policeman shot the criminal dead."
('Ge,' in this case, is not a pre-verbal marker but a past tense marker that's replaced any time a pre-verbal marker comes into play.) The fiendishly difficult part of all this for non-native speakers is that it's infuriatingly inconsistent. For example, 'er' often means completion with a distinctive change of meaning ('trinken' to drink, 'ertrinken' to drown).
You should notice that this is very different from the Fennic object marking, which is mandatory inflectional agglutination. Prefixing verbs is optional derivational agglutination. You cannot ignore or fail to use the Fennic object cases, lest you produce ungrammatical sentences; but, you can ignore a preverbal marker and use alternative expressions. Furthermore, although occasionally the Fennic telic entity is also lexical (telic and atelic are different words), most of the time it's purely grammatical. Whereas, preverbal markers are always lexical: you get a completely different verb.
The German verbal markers would be comparable to Finnish frequentatives, momentanes, and causatives, which are verbal markers not really expressible in full in English. For example, causative räjäyttää is "to make something explode", is a different verb from the momentane räjähtää "to suddenly explode by itself", just like German schossen is a different verb than erschossen, even if they are grammatically marked with verbal infixes -autta- "to force something" or -ahta- "suddenly (anticausative)".
The difference with perfectivity would be this: Löin naulat (telic) vs. Löin nauloja (atelic), "I hammered the nails". The telic meaning would suggest that the work is done, irrespective of such oppositions if the work was to hammer the nails half-way or whole way in, or leave some without hammering, etc., as indicated by perfectivity.
As for translation - it just doesn't happen systematically. To translate telicity requires a complete reassessment of the meaning of the sentence. Similarly, to translate from a language without this distinction, care needs to be taken that telicity is selected correctly. Let's take a Microsoft translation for a bad example:
Original English: Looking in digital media folders first, then everywhere else.
Microsoft translation: Etsitään digitaalisen median kansiot ensin, sitten kaikkialta muualta. "Finding digital media's folders first, then from everywhere else."
Intended translation: Etsitään digitaalisen median kansioita ensin, sitten kaikkialta muualta. "Searching about in the digital media folders first, the from everywhere else."
As you can see, the meaning is broken in the translation. The translator ignored the word "in" in the original English. He then translated the following words word-by-word, which produces the accusative
case. Thus, the word "folders", "kansiot", lands on the accusative (telic) meaning. This produces the meaning that the searching is being completed, as its intended result is finding. The meaning changes into "Finding the folders." Should he have selected the partitive case
, it'd have produced the meaning "Searching about the folders, yet unsuccessfully", or in English grammar, "Searching around in the digital media folders first."
Exercises for the reader...
- A mosquito stung me.
- Will you help us out of this?
- Node that, please.
- I will write a book.
To view the small print better, enlarge the font. (In Mozilla, this is Ctrl-+ to enlarge and Ctrl-- to diminish.)
1. Atelic: mosquito bites will not kill you, and even if they did, that is not indicated in the sentence. 2. Telic, the result "out of this" is indicated. 3. Telic; "to node (a fact)" governs only the accusative case, as noding is posting a web site. Posting a web site cannot be "in progress"; before or during the transmission of the data to the server it has not "been noded" as yet, and after the transmission is complete, it has "been noded". 4. Telic, result is indicated and in the future.
The small print:
- The translation "He drank up the coffee" doesn't catch the telicity in "Hän joi veden" (He drank coffee-telic). The problem is that "to drink up" is a perfective, not a telic verb, i.e. it only indicates the action has been finished, not if it was successful, even if in this case these two coincide. If telicity wasn't independent from the perfective aspect, there wouldn't be an independent expression for "drink up" in Finnish. But, there is: "juoda .. pois" - "to drink .. away". This is a perfective verb, which can assume either telicity. It is distinct from telicity, and thus can be superpositioned with it, namely:
1. Hän joi (pois) kahvin. (telic - accusative) "He drank up the coffee."
2. Hän joi pois kahvia. (atelic - partitive) "He was drinking the coffee up."
In example 1, the word pois "away" can be omitted, because its only function is emphasis, as the perfectivity is included in the indicated telicity. Compare this to English: should we omit the word "up" in "He drank up the coffee", we get the sentence "He drank the coffee", with no indication of completeness. In this way, it is shown that telicity is not only distinct, but can actually supersede perfectivity, leaving it only an emphatical function.
Example 2 is better mirrored with this: "He was drinking up the coffee, which was yet unfinished by him." If telicity was the same as the perfective aspect, that expression should have the same meaning as "He drank coffee." However, as you can see, "He drank coffee" conveys much less information than "He was drinking up the coffee, which was yet unfinished by him." The latter explicitly denotes if the coffee is finished. Thus, telicity is distinct.
Also, we haven't even discussed the fact that "successfully completed as intended", or telic, is not the same as "stopped", or perfective.
Another point is the word "the". It only indicates that we know what coffee he drinks, not that he finishes it. The word "the" can't totally corner the idea of completion and has no actual telic meaning without context. If we reconsider the meaning of the accusative, the accusative can be replaced with "that". This is illustrated by the fact that when expanded as "He drank the coffee while reading the newspaper", both telic and atelic translations are valid, but differentiate the telicity English cannot explicitly denote. The sentences are:
Hän joi kahvin lukiessaan lehteä. — this telic one indicates he finished.
Hän joi kahvia lukiessaan lehteä. — atelic; unknown if finished.
The partitive ending is basically -ta, but this is not the whole story. Vowel harmony requires A to morph into Ä, if the word does not start with a back vowel (A, O, U). The 't' usually elides, if you add the ending next to a vowel, as in above "kahvia". A consequence is that if word already ends in -ta, it's -taa, e.g. rantaa. Also, there are exceptions with a geminate 'tt' such as haketta "wood chippings (part.)", or venettä "boat (part.)" A consonant, or a if there was one historically, causes the geminate to occur; venet was the 19th century form of modern vene, the consonant surviving as venes in Pohjanmaa dialect.
The accusative ending in Proto-Uralic is -m, the same as in Latin. This has developed into -n; Finnish elides it in some positions, such as with the imperative. For example, "eat the fish": *"syö kalam" → "syö kala", but "I eat the fish": *"syön kalam" → "syön kalan". Unfortunately, the former (no ending) is identical to a nominative ("uninflected form"), and the latter (ending -n) is identical to a genitive (someone's).
- Ackerman, Farrell; Moore, John. Telic Entity. URL: http://csli-publications.stanford.edu/LFG/4/ackerman-moore/lfg99-ackerman-moore.html