I'm not familiar with the concept of phase. But judging from what you have quoted, it seems to me that phase refers to an objective characterisation of events, whereas aspect refers to a choice that a speaker makes in how to refer to events. A journey will have a beginning, a middle, an end, an entirety, and other objectively characterised properties and phases. When we refer to the journey we have a choice whether to talk about it as a completed unit (perfective), about it as a continuing process (imperfective).

This link talks about aspect and phases, but with a different distinction than what you describe: blclab.org/wp-content/uploads/2013/02/nllt00.pdf

@ColinFine Thank you for your answer. For lack of space, I respond in two comments. Ok for perfective and imperfective, but what about the inchoative aspect vs. the starting phase (for instance)?

@ColinFine In the Wikipedia article, it reads: "Inchoative aspect (abbreviated inch or incho) is a grammatical aspect, referring to the beginning of a state. (...) The English language can approximate the inchoative aspect through the verbs "to become" or "to get" combined with an adjective." To me, "to become" + something sounds objective. And also, it seems very similar to Talmy explanation of phase : "Phase notions can also appear as the sole meaning of a verb without the incorporation of further semantic referents, as in English start, stop, finish." [the last 3 verbs should be italicized]

Yes, and what you quote says English can approximjate the inchoative aspect, not English has an inchoative aspect. He is becoming angry is a different claim from he is angry, with different truth values, so the choice is not purely an aspectual choice.

@Vegawatcher Thank you for the article, it is indeed enlightening, especially on this notion of aspect which, the authors demonstrate, is not very clear. Indeed, the "phase" they are referring to seems not to be related with the "phase" I was asking about here. Thank you again

@ColinFine If we take the English progressive aspect "be + -ing" as an example of aspect: He swims and He is swimming have the same truth value? I find it difficult to agree on this point, because 'to have the habit to swim' and 'to be swimming now' are as much different, to me, as 'to become angry' and 'to be angry'.

You're right. Theapproach I was arguing for doesn't work for all cases. Aspect is difficult in English, because it is confounded with other things. FOr example, in narrative, goes/is going is a simple/progressive (or perfective/imperfective) distinction. In normal (non-narrative) speech, this distinction is not available to us grammatically for most virbs, and instead we have the distinction between a true prsent and a timeless or habitual. kI agree that that is reasonably referred to as "aspect", but it's a different distinction from the one in narrative.

@ColinFine I guess it would be easier if I knew a slavic language! I edited my post by trying to show the difficulty from the point of view of a language, Chinese, which does have some aspect markers, and by trying to show in more details how the theoretical explanation of what is phase compared to aspect can be challenging. I also provided my own interpretation

Perfective/imperfective opposition is not grammaticalized/lexicalized in Serbian, and I'd guess the same is true of other Slavic languages. Unlike English, Serbian doesn't have a dedicated grammatical form for progressive meaning, or any other kind of verb aspect. What is lexicalized is the telicity of the verb - each verb is clearly telic or atelic (which of course contrasts with English where such opposition doesn't exist).

Of course, by virtue of being telic, the verb will by necessity be perfective, but it doesn't mean that the expression of perfectivity is restricted to these verbs. The verb form that is in popular literature erroneously labeled "imperfective" , is best named as "unmarked", since it can be used to express any aspectual meaning, including the perfective one. You might want to check my answer here linguistics.stackexchange.com/questions/25742/…

Confusing the concepts of perfectivity and telicity and sticking to the obviously untenable claims of the lexicalized perfective/imperfective opposition in Slavic languages, led authors to logically inconsistent analyses. Here is a quote from a work on aspect escholarship.mcgill.ca/downloads/m326m213f.pdf : "....The apparent counterexamples that have been much discussed in the literature have to do with so-called ‘telic’ readings of imperfective verbs - that is to say with the use of imperfective forms in the situations that are perceived as completed at the time of speech (ST):

(63) Adapted from Forsyth (1970), Schoorlemmer (1995) and Borik (2002) a. Ja ne pojdu v kafe. Ja (uže) poela/ela . (I'm not a Russian speaker, but the languages are close enough, the same illustration could have been taken from Serbian). Instead of recognizing the fact that the perfective/imperfective division of Slavic verbs cannot be supported by the facts of language, the authors talk about "telicity of imperfective verbs" and provide "counterexamples", which are supposed (erroneously) to be thought of as exceptions to the unquestionable dichotomy.

@user145148 Hi! By "I guess it would be easier if I knew a slavic language!", I meant that, since slavic languages are aspect prominent languages, it would be maybe easier, knowing one of those languages, to understand the difference between aspect and phase. For instance, maybe there is a very clear semantic difference in those languages between a lexical verb of action marked with the inchoative aspect and phrases containing a verb with the meaning 'start' followed by a lexical verb of action

Hi starckman! Unfortunately, I can't provide a competent answer on the nature of "inchoative" aspect in any of the languages, including my mother tongue of course. I just argued against the general understanding of the Slavic aspect. Slavic languages can hardly be thought of as "aspect prominent". English is far more so, as it obviously has a grammaticalized "perfective/progressive" opposition, as far as event verbs/single occurrence situations are concerned.

Those "counterexamples" that authors use to point to "telic imperfective verbs", as a sort of anomaly in Slavic languages, are anything but exceptions. I wouldn't have to work hard to come up with any number of examples of such verbs - using those so called "imperfective verbs" to express perfective situations is a natural property of the Serbian language. By saying "Ja sam jela" I'm not making any sort of claim towards the telicity of the event of eating - but I certainly make a claim about the perfectivity of the event - it is a perfective event. Again, confusing telicity with perfectivity

..is bound to lead to a number of illogical conclusions. On the other hand, by saying "Pojela sam..." I make an explicit claim regarding the telicity of the event - the situation reached the culmination point. The point is that the distinction jesti/pojesti (or the analogue in Russian) does not concern the perfectivity vs imperfectivity of these verbs - it concern their telicity. Again, the telic/atelic dichotomy is in-built in the Serbian language, the telic part of the opposition pointing to the culmination of the event in question. There is an intricate mesh of morphemes and processes that

change the telicity of the verb. The naturally atelic verbs are turned into telic ones by a number of "telicizing" prefixes ( "pojesti" is best understood as the marked part of the opposition, formed by adding the prefix "po" to the unmarked, atelic "jesti"). There are verbs which are naturally telic, and the reverse process is in order if we want to make them atelic (roughly speaking, in such case, the telic form will be tweaked by a suffix of a sort (e.g prodati (sell) becomes "prodavati" (to be selling)).

The sentence "Ja sam jela" (in the given context translates perfectly as English "I have eaten" (existential perfect meaning)), perfectly illustrates the complexity of the problems that "perfective/imperfective" dichotomy of Slavic verbs leads to. The author of the cited work speaks of ".. ‘telic’ readings of imperfective verbs " discussing this and similar examples. The problem is that "jela" in this sentence is neither telic nor imperfective, it is the other way around.

Just like the perfectly corresponding English translation "..have eaten" in "I have eaten", "jela" in "Ja sam jela" refers to a PERFECTIVE and ATELIC situation. The author clearly misunderstands and confuses the concepts of "completion" and "pastness", (cit. - ..that is to say with the use of imperfective forms in the situations that are perceived as COMPLETED at the time of speech) and it is a huge mistake. At the time of speech the event of eating is obviously finished, but not necessarily completed.

@user145148 If I understand you well, telicity must refer to the completion of an action, in other words to the attainment of the goal that this action is supposed to lead to (e.g. he built the house), and perfectivity to the mere end of an action, regardless whether it has been completed or not (he ate a cake). And Serbian expresses obligatorily telicity, not perfectivity.

Both "I have eaten" and "I have eaten a cake" are perfective events but only the latter is telic. "I have eaten" , or in Serbian "Ja sam jela" presents the situation of eating as a single whole, so it is perfective. The situation is also atelic - (a)telicity is another distinct dimension of the event, to be distinguished from the concept of perfectivity. Depending on our communication needs, atelic situations can be presented perfectively or imperfectively. Telic events can receive only perfective interpretation, simply because their culmination point is in-built - you can think of the event

..only as one single whole, if your description of the situation contains a clear indication of the culmination point. Serbian verbs are either telic or atelic: "jesti" is an atelic verb, "pojesti" is telic. Each verb can only be one or the other. This is not true of English verbs. The verbs eat or build that you mention, can head either telic or atelic situations, they are not either in themselves. "Build houses" is atelic, "build a house" is telic, "eat a cake" is telic, etc. Now, when it comes to perfectivity it is where English has an apparatus to express this dimension of verbs:

non-recurrent event verbs in the simple past tense are perfective, and those same verbs in the ing form are progressive and thus imperfective. There is no such distinction in the Serbian language: "Ručala sam u restoranu" can equally mean either "I was eating my lunch at a restaurant" or "I ate my lunch at a restaurant" and there is no way we can know which is the intended meaning without getting into the context. In my answer in another thread, I provided quotes from the Cambridge Grammar and another work on aspect, and the authors take a completely different position on this issue.

Here's from CGEL: "In languages such as Russian there are distinct verb-forms whose basic meanings correspond closely to these two aspectualities, and these languages are therefore said to have perfective and imperfective aspect. English, of course, is not such a language: the simple present and preterite can both be used either perfectively or imperfectively." I claim the opposite: English does have distinct verb forms which correspond to non-progressive and progressive aspects (the specific class of verbs). I doubt Russian verbs can be divided along that line, and I know that Serbian can't.