Canonical Post #2: What is the perfect, and how should I use it?

Question

This is a Canonical Post, intended as a reference and resource for both Questioners and Answerers.

---

The English “perfect” is deeply puzzling for learners. Nearly one Question in every twenty here asks about perfect constructions, and every Answer seems to raise new Questions. Even very advanced learners often misuse the perfect, or fail to use it when they should.

If it makes you feel better, the experts are baffled, too. Grammarians and linguists have been quarreling about the perfect for more than two hundred years. There are several large books on the subject, and important papers are published every year. It is only in the last ten years or so that a consensus has started to emerge.

Moreover use varies, especially in speech. Some uses are statistically more common in the UK than the US, and in many contexts speakers make no very clear distinction between, say, a present perfect and a simple past.

So there are no cut-and-dried rules you can always apply in every situation. The best I can offer is rules of thumb. These will work in nine cases out of ten — and nobody has yet figured out the rules for the tenth case! … I’ll mark these with this sign: ☛. Here’s the first one:

☛ Follow formal use of the perfect
This may appear to contradict what many of you have been taught, which is to keep your language colloquial and ‘everyday’. But with English perfects, formal use is everyday. The standard forms are used and understood everywhere, and will not mark your speech as pedantic or unnatural, even in very casual conversation. It will be simpler for you to learn just one pattern and use it all the time.

My rules of thumb will always reflect standard formal use. I will take little notice of colloquial use, except to point out non-standard (but colloquially acceptable) uses which might confuse you. With examples I’ll use a leading asterisk, ^∗, to mark utterances which are regarded as non- or sub-standard in the formal register.

^∗ I haven’t ate yet.

---

Because this is a very large topic, I’ve split it into separate Questions. Here are links to them, with the “Short Answer” rules of thumb for each:

• 1. How do I construct a perfect?

☛ English perfects have two components: a form of the auxiliary verb HAVE on the left (the ‘HAVE’ piece), and a verbstring headed by a past participle (P_^A·P_^PL) on the right (the ‘VERB’) piece).

This section includes a discussion of the most common errors learners make in constructing perfects.

• 2. How do I recognize a perfect in context?

☛ A perfect construction is a form of HAVE followed by a past participle, with nothing coming between them but adverbs, adverbials or negations.

This section includes discussion of (2.1) syntactic factors which make recognizing a perfect construction difficult, (2.2) constructions which look like perfects but aren’t, and (2.3) constructions with modal verbs which employ the perfect construction in a non-perfect sense.

• 3. What does the perfect mean?

☛ The perfect introduces a prior eventuality which in some sense constitutes a current state. But it is up to the hearer to infer the nature of that state.

This is a long and complicated section, so I have divided it into two parts:

• 3.1 grammatical meaning of the perfect, which is conferred by the construction itself

☛The VERB piece presents an eventuality located before the time which is being spoken about.

☛The HAVE piece presents a state which is current at the time which is being spoken about.

☛The perfect cannot be used to express narrative sequence.

• 3.2 pragmatic meaning of the perfect, which is inferred by the hearer/reader from the context in which the construction is used.

☛ The ‘standard framework’ describes what the perfect means, distinguishing three meanings the perfect may express: continuative (“has been since...”), resultative (“has brought about ...”) and existential (“been there, done that”).

☛ Recent studies look at how the perfect means and suggest that meaning is not expressed by perfect constructions but inferred by hearers/readers from both the prior eventuality introduced and the larger discourse context.

• 4. When and how should I use the perfect?

☛ “Don’t use the perfect unless you need it.”

☛ Use perfect constructions to introduce prior eventualities as context for the current discussion.

• 5. ELL questions about the perfect

This section links many ELL questions and answers which may address your immediate concern more directly. It is divided into questions about

• constructing a perfect correctly
• distinguishing perfect constructions
• perfect tense
• perfect aspect
• modal, irrealis and conditional perfects
• the kinds of perfect meaning
• choosing between perfect constructions and simple present or past constructions
• coordinating the use of perfect constructions with clauses using other verb constructions
• the use of perfect constructions with time expressions

A note on notation

Occasionally the names of verb constructions will appear abbreviated, as follows:

• PR: present
• PA: past
• FU: future
• SI: simple
• PRG: progressive
• PF: perfect
• PSV: passive
• PPL: participle
• INF: infinitive

These abbreviations may be combined, thus: P_^A·P_^F·P_^RG, meaning a ‘past perfect progressive construction’.

A word in uppercase italics, like HAVE, means ‘any appropriate form’ of the word, and the name of a word class (part of speech) in uppercase italics, like VERB or MODAL, means ‘any word of that class’. The superscript abbreviation ^X_X of a verb form or construction after one of these means ‘the XX form of this word or word class’. For instance,

VERB^P_A·P_PL means ‘the past participle form of whatever verb you are using’.

NEXT: 1. How do I construct a perfect?

Answer 1

2. How do I recognize a Perfect in context?

SHORT ANSWER:

☛ A perfect construction is a form of HAVE followed by a past participle, with nothing coming between them but adverbs or adverbials.

LONG ANSWER:
In theory perfect constructions should be easy to identify: they’re always marked by a form of HAVE followed by the participle of the next verb. But in actual use it’s not so easy.

2.1 Disguised perfects

First, there are constructions which make it difficult to detect a perfect construction:

• Intrusions  Adverbs and negators may be placed between the HAVE form and the past participle which completes the construction:

  John has recently written the manufacturer about this problem.
  She had never seen such a beautiful sight.

  There may be several intruding adverbials:

  He has deliberately, maliciously and quite illegally denied my application.

  Intruding adverbial phrases may be v e r y   l o n g—and there’s no guarantee that writers will do you the courtesy of bracketing them in commas.

  He had without stopping to think whether there was any real likelihood that it would effect an actual improvement in his position spoken to Prof. Sartorius about his grade.

  That’s not good writing. But I’m afraid that a great deal of what you read is going to be not-good-writing. You will encounter this sort of thing frequently, so you need to watch out for it.

  Additionally, if you're reading older poetry, you may sometimes find direct objects, noun complements, etc., between the HAVE form and the past participle, usually for metrical reasons.

• Questions, and other cases of subject–auxiliary inversion  Ordinarily, only adverbials can intrude between the HAVE form and its past participle complement. But there are various situations in English where an auxiliary verb switches places with the subject; when that happens in a perfect construction, the result is usually that the subject intrudes between the form of HAVE and the past participle.

  The most common such situation is in questions:

  Has Prof. Sartorius denied my application?

  Other such situations are more common in literature than in speech:

  No sooner had that problem been solved than a new one arose.
  Had we known then what we know now, we would not have agreed to the undertaking.

• Omissions  ‘Ellipsis’—omitting repeated or implied words—may result in the past participle of a perfect construction left without an adjacent HAVE form:

  By purchasing this product, you certify that you have read and accepted the Privacy Policy and Terms and Conditions... [link]
  The economy has neither completely recovered from the global recession which struck in 2008 nor remained permanently stuck in a protracted depression.
  [colloquial] You been there yet?

  Conversely, ellipsis may result in the HAVE form appearing without an adjacent past participle:

  I haven't seen it, but he has.
  You've been there, haven't you?

• Adjectival participles  Note that in one of the examples above, there are also past participles which are not part of the perfect construction—stuck and protracted. These are employed as adjectives—stuck as an adjective complement and protracted as an attributive adjective. That’s something else you have to watch out for.

2.2 Sham perfects

Second, there are constructions which look like perfects but aren’t, because HAVE is being used with a past participle, not as an auxiliary but as a lexical verb.

• Causative HAVE  In the first of these constructions, HAVE carries a causative sense: the Subject of the sentence causes something to happen. It is used with a subordinate clause in the passive voice from which the BE form has been deleted.

  Sam is having his house painted. = Sam is causing his house to be painted.

  Col. Sartorius had the mission carried out by Lt. Trench. = Col. Sartorius caused the mission to be carried out by Lt. Trench.

• Resultative HAVE  In the second construction, HAVE is used approximately in the ordinary sense of “hold” or “possess”, and the past participle acts as an adjective modifying the preceding noun or noun phrase. The construction is called ‘resultative’ because the past participle describes the state of the noun which results from the action of its verb.

  We can’t act until we have this question settled. = ... until the question is before us in a settled state.

  Now that she has her kids graduated she has a lot more time to work with us. = … the kids are in a graduated state.

  Ordinarily it’s easy to distinguish these two constructions from perfect constructions: a noun phrase intrudes between the HAVE form and the past participle. But if the noun phrase is moved the difference is less obvious:

  Question 9604 (user Listenever)

  “I’ve come to bring Harry to his aunt and uncle. They’re the only family he has left now.” — Harry Potter and the Sorcerer’s Stone by J.K. Rowling.

  OALD says leave somebody is to have family remaining after your death (OALD, leave #10). But the case’s subject is not Harry’s parents but Harry. Can Harry, who lived, be the subject of ‘leave’?

  User Listenever, perfectly reasonably, takes the underlying construction to be a perfect, he has left (some) family now. But what is actually going on is that the underlying construction is resultative: he has (some) family left now = he still has some family. This has been ‘transformed’ into a relative expression, leaving an ‘empty’ space between the HAVE form and the past participle: …family (which) he has ∅ left now.

• HAVE got  As I mentioned earlier, the past participle of get in North American (US and Canadian) use is gotten, so HAVE got cannot be understood as a perfect construction there. It is simply a colloquial and somewhat more emphatic variant of HAVE.

  I’ve gotten four steaks for our dinner. = I have obtained four steaks for our dinner.
  I’ve got four steaks for our dinner. = I have four steaks for our dinner.

  But got is the ordinary past participle of get in other Englishes, and this can cause some confusion. HAVE got in the North American sense is spreading to those other speech communities, so if an Englishman or Australian says “I’ve got four steaks” he may mean either I have or I have obtained. You have to figure out what they mean from context.

  Note that HAVE got = HAVE is used only with HAVE in the simple present. With any other form or construction (had got, having got, to have got, MODAL + have got), HAVE got is a perfect

2.3 Modal perfects and sham perfects

Finally, there is the very quirky use of HAVE + past participle with modal verbs and in irrealis (‘unreal’) expressions.

The English ‘full modals’—can/could, may/might, shall/should, will/would—and irrealis expressions are (as you probably know) even more complicated than the perfect; so when these are combined with the perfect you have to expect especially difficult constructions. I haven’t got space here to go into all the details, but there’s one thing I have to warn you about: in these situations the HAVE + VERB^P_A·P_PL construction is not always a perfect.

Sometimes it is. Will have VERB^{P _A·P_PL} for instance, is the ordinary way of expressing future perfect, a future state grounded in prior eventualities. Likewise, may have VERB ^P_A·P_PL can be used to express the possibility of a present state grounded in prior eventualities.

Bob says John will have fixed the program by tomorrow. = Bob predicts that tomorrow we will be able to say ‘John has fixed the program’.

Bob thinks John may have fixed the program already. = Bob believes it is possible that John has fixed the program already.

The simple past forms of these auxiliaries, would and might, may be employed to express futurity and possibility in the past.

Last Wednesday Bob said John would have fixed the program by the next day. = Bob predicted last Wednesday that on Thursday we would be able to say ‘John has fixed the program’.

Last Wednesday Bob thought John might have fixed the program already. = Bob believed last Wednesday that it was possible that John had already fixed the program.

But there’s an ambiguity in those last two sentences. It arises because the simple past forms are also used to express irrealis mode—’unreality’—in the present.

If he had_IRREALIS the right software, John would fix_IRREALIS the program right now.

If you want to express that in the past tense, you have a problem—you’ve already used up your past forms of these verbs! To get around this, the language employs the perfect construction as a ‘past marker’.

If he had had_IRREALIS the right software, John would have fixed_IRREALIS the program right then.

In this case, had had and would have fixed look like perfects, but they aren’t interpreted as perfects. What they ‘mean’ is the ‘irrealis simple past’ of ‘irrealis simple present’ had and would fix.

NEXT: 3. What does the perfect mean? - 3.1 grammatical meaning of the perfect

Answer 2

3. What does the perfect mean?

SHORT ANSWER:

☛ The perfect introduces a prior eventuality which in some sense constitutes a current state. But it is up to the hearer to infer the nature of that state.

LONG ANSWER:
If you have been relying on concepts like ‘completed action’, ‘anterior event’, ‘indefinite past’, ‘current relevance’, or ‘incomplete action’ to understand perfect constructions, you need to know that none of these formulas are reliable. Sometimes one works, sometimes another. They’re echoes of many theories put forward in the course of a prolonged debate about what the perfect means.

Grammarians have not yet arrived at a final theory. There has, however, been very considerable progress, and a consensus appears to be emerging. Today it is generally agreed that perfect ‘meaning’ has two components:

• a grammatical meaning, which is conferred by the construction itself, and
• a pragmatic meaning, which is inferred by the hearer/reader from the context in which the construction is used.

I’ll discuss grammatical meaning in this section, and pragmatic meaning in the next. I’m afraid the grammatical piece is fairly technical; but if you find your eyes glazing over, skip it and go to the section on pragmatic meaning. There’s a summary there, and you can always come back to the technical nuts-and-bolts later.

3.1 Grammatical meaning

We’ve seen that the perfect construction has two parts: a form of HAVE on the left and a lexical VERB on the right. In actual sentences, however, what’s on the left may be preceded by a modal, and what’s on the right is an entire Verb Phrase, including auxiliaries and complements.

I’m going to ignore the technicalities, mostly, and refer to the two parts as the HAVE piece and the VERB piece.

It may be helpful in what follows to think of the HAVE piece and the VERB piece as if they contributed different parts of the meaning. This is not strictly true—but it’s a handy way of thinking about it, and it does reflect some syntactic facts.

---

English verb constructions typically ‘encode’ several properties: modality, voice, viewpoint aspect, tense and lexical aspect. The perfect construction, however, is a complicated device, and we have to distinguish those properties which belong to the construction itself from those which belong to entities lying outside the construction or embedded inside it. Only tense and lexical aspect are directly involved in the perfect itself—the pieces in red in this graphic:

3.1.1 Voice and Viewpoint aspect

The VERB piece, on the right, can be cast in progressive or passive form, or both; but this does not affect the interpretation of the perfect construction itself. Consider the sentence I have expressed this verb in the perfect .

• Only the VERB piece, express, may be transformed into the passive voice, not the HAVE piece:

  ^OKThis verb has been expressed in the perfect passive BUT NOT
  ^∗  This verb is had expressed in the perfect passive.

• Similarly, only the VERB piece may be transformed into the progressive aspect, not the HAVE piece.

  ^OKI have been expressing this verb in the progressive perfect . BUT NOT
  ^∗  I am having expressed this verb in the progressive perfect .

3.1.2 Modality

Modality is expressed with a modal verb heading the perfect construction. The primary tense is ‘moved’ onto the modal which now heads the construction, and HAVE takes the infinitive form.

But as I explained at the end of the last section, modal verbs and irrealis (‘unreal’) expressions also use what looks like a perfect construction to mark tense. To keep things simple I’m going to deal only with the ordinary indicative meanings of perfect constructions here.

That leaves tense and lexical aspect. The tag-wiki articles on these—tense and aspect—have fuller, more technical descriptions, which I’ll just summarize here.

3.1.3 Perfect Tense

There are actually two time relationships at play in perfect constructions. Tense1, which is ‘encoded on’ the HAVE piece, relates Speech Time (ST), the time when we speak or write the sentence, to Reference Time (RT), the time which we are speaking about. Tense1 may be simple present (have,has) or past, or, with a modal, future (will have), or it may be unmarked for tense (non-finite). This is the primary tense in the sentence; that’s why you are not allowed to use past adverbials with the present perfect:

^∗  John has finished the report yesterday.

Here has is present-tense; yesterday clashes with this just as badly as in John is yesterday.

Similarly, the next sentence jars because present-tense has clashes with the fact that its subject has been dead for centuries:

^∗  Gutenberg has invented printing.

Tense2 is attributed to the VERB piece. It always locates that piece before the HAVE piece, and therefore before RT.

☛The VERB piece presents an eventuality located before the time which is being spoken about.

Be careful: This not quite the same thing as what many learners believe:

“The perfect is used to say that one action occurred before another.”

Although perfect constructions do say that one eventuality precedes another; these constructions are not used in order to say that one eventuality precedes another, as we will see next.

3.1.4 Perfect Aspect

Regardless of the lexical aspect of the VERB piece, the construction is syntactically presented as a state which ‘belongs’ to the Subject of the sentence.

Let’s take the sentence John had written three novels—an accomplishment verb in the past perfect—and subject the result to the three stativity tests in the ‘aspect’ tag-wiki:

• States may not be employed with the progressive construction:

  ^∗  John was having written three novels. –This is not idiomatic English.

• In a main clause modified by a when clause, states are understood to overlap—to precede and to continue during—what is described in the when clause.

    When John was 25 he had written three novels. —John’s novels were complete by some time in his 25th year, and of course they continued to be complete afterwards.

• States cannot serve as the complement in Wh-cleft constructions.

  ^∗  What John did was have written three novels. –This is not idiomatic English.

The last test is particularly interesting. Note that we can make the Wh-cleft work by moving the HAVE piece out of the complement onto the main verb:

^OKWhat John had done was write three novels.

That confirms that the VERB piece, the eventuality located before RT, is syntactically subordinate. The eventuality which the construction expresses is a state, which is located at RT:

☛The HAVE piece presents a state which is current at the time which is being spoken about.

This has the important consequence that

☛The perfect cannot be used to express narrative sequence.

The perfect presents a state, not an event, and states do not ‘occur’ but ‘endure’. In consequence, consecutive perfects express states which are understood as simultaneous or overlapping, not consecutive.

A classic example is the indictment of George III in the US Declaration of Independence:

He has refused his assent to laws, the most wholesome and necessary for the public good.
He has forbidden his governors to pass laws of immediate and pressing importance ...
He has refused to pass other laws ...
He has called together legislative bodies at places unusual ...
He has dissolved representative houses repeatedly ...

... and so forth: seventeen present perfects, interrupted by a single present progressive (which is also a stative construction). Jefferson does not say that ”The King refused and then he forbade and then he called ...”, he describes “a history of repeated injuries and usurpations” which constitute the present justification for the Colonies’ action.

NEXT: 3. What does the perfect mean? - 3.2 pragmatic meaning of the perfect

Answer 3

3. What does the perfect mean? (part 2)

3.2 Pragmatic meaning

To summarize grammatical meaning: Perfect constructions have two pieces, a form of HAVE on the left and a VERB with its auxiliaries and complements on the right.

☛The VERB piece presents an eventuality located before the time which is being spoken about.
☛The HAVE piece presents a state which is current at the time which is being spoken about.

But what is that ‘perfect state’?

3.2.1 The standard framework

In 1971 James McCawley¹ described three fundamental uses of the English present perfect (the same uses are also present in past and future perfect constructions). With minor modifications, and under a variety of names, those three uses are to this day the standard framework for discussing the meaning of the perfect.

1. the continuative or universal perfect (perfect of persistent situation) indicates a prior state which continues to be in effect right up to Reference Time.

   John has lived in Paris since 2009.

   The VERB in this perfect must be stative, or recategorized as stative—for instance, by being cast in the progressive:

   Amy has been running for the last hour.

2. the resultative or stative perfect (perfect of result) indicates at least one prior event whose result state is still current at Reference Time.

   Bob can’t be here tonight—he has caught the flu.

   The VERB in this perfect must be telic—must give rise to a new state.

3. The existential or experiential perfect indicates the existence of past events: it asserts that at least one instance of the eventuality named by VERB occurs prior to Reference Time.

   Michaelis has written several papers about tense and aspect.

   The VERB in this perfect may bear any lexical aspect.

(McCawley also proposed a ’Hot News’ perfect (perfect of recent past) announcing a recent event as new information.

John has won the Nobel Prize!

But in 1981 McCawley withdrew this as a distinct category, and it is now generally agreed that uses of this sort are special instances of existential or resultative perfects.)

The McCawley categories are an excellent starting point for a learner. They describe all the uses you’re going to encounter, and help you interpret them. They distinguish meanings and explain when and why the various formulas you hear apply. For instance, questions here suggest that different learners have heard that the perfect expresses both ‘completed’ and ‘incomplete’ action. As we see above, these formulas apply to different uses of the perfect: resultatives and existentials assume completed actions, while the continuative assumes a continuing action.

☛ The standard framework describes what the perfect means, distinguishing three meanings the perfect may express— continuative (“has been since...”), resultative (“has brought about ...”) and existential (>“been there, done that”).

But the McCawley categories are only a starting point.

• To begin with—and this is a constant problem for learners—the categories don’t explain when or why a speaker would employ an existential present perfect rather than a simple past. Both appear to express the ‘existence’ of a prior event:

  John has written five papers about the present perfect.
  John wrote five papers about the present perfect.

  What does has written say that wrote doesn’t? To put it another way, in the context of our understanding of the grammatical meaning of the perfect, What is the state expressed by John has written five papers?.

• Furthermore, it’s not always clear what category a specific perfect belongs to. A sentence like “John has lived in Paris” may be read as either continuative or existential:

  CONTINUATIVE: John has lived in Paris ...since 2009.
  EXISTENTIAL: John has lived in Paris ...on three different occasions.

  And in some cases an existential may have what appear to be resultative implications:

  ^? RESULTATIVE: John has lived in Paris → so he probably speaks pretty good French.

  The technical literature now finds it necessary to distinguish two sorts of ‘resultative’. A result which is a necessary consequence of the eventuality—specifically, a result which is the end state of a telic verb—is characterized as ‘strong’ or ‘entailed’:

  Bob has caught the flu → Bob has the flu.

  A result which is not a necessary consequence is characterized as ‘weak’ or ‘implicated’:

  Bob has caught the flu → Bob isn’t at work today.

3.2.2 A post-modern framework

What is clear from this is that is not the perfect constructions themselves which express the McCawley meanings, but the contexts within which the constructions are used.

Consequently, the most interesting recent studies of the perfect have focused on the pragmatics of the construction—how its meaning is derived from interaction with the context. This is still very much a work-in-progress; but the approach in a recent paper by Nishiyama and Koenig² strikes me as particularly helpful.

• These authors shift the focus from the way in which meaning is encoded to the way in which meaning is decoded. They claim that the salient characteristic of the perfect construction is an “underspecified” perfect state—that is, a perfect construction establishes that there is a relationship between a prior eventuality and a current state of affairs; but it does not explicitly define the relationship or the current state. That definition is left to the hearer to infer.

• What supports this inferential understanding of perfect constructions is a fairly well-established ‘rule’ of discourse, Levinson’s principle of informativeness:

  Speakers are called on to “Say no more than you must”, choosing a less informative utterance when a more informative one is available.
  Hearers are called on to “Understand as much as you can”, finding the most specific interpretation they think the speaker intended.

• The perfect uses described in the standard framework reflect different sorts of inference applied to the task of defining the perfect state.

• A small corpus study of 605³ real-life uses of the perfect supports this conclusion.

  • In most cases the necessary inference was almost trivially simple. More than 80% were governed by a ‘presumption-of-persistence’ rule: hearers assume that “states that they have been told about persist until they are given evidence to the contrary”. These cases all yielded what the McCawley framework classifies as continuative or entailed resultative uses.

  • The remaining cases, which yielded implicated resultative or existential uses, required more complex inferences based on shared knowledge about the immediate situation and ‘commonsense’ assumptions about the world. In almost every case, the relevant immediate facts were expressed in the immediate context, either the preceding or the following sentence.

An interesting conclusion to be drawn from this framework is that although perfect constructions have ‘default’ interpretations, these are not entailments but implicatures. The default interpretations may be ignored or explicitly negated.

☛ Recent studies look at how rather than what the perfect means and suggest that meaning is not expressed by perfect constructions but inferred by hearers/readers from both the prior eventuality introduced and the larger discourse context.

NEXT: 4. When and How should I use the perfect?

---

¹ James D. McCawley, “Tense and time reference in English”, in Langendoen & Fillmore, Studies in linguistic semantics, 1971.

² Atsuko Nishiyama and Jean-Pierre Koenig, “What is a perfect state?”, Language 86, 3, 2010.

³ 605 perfects drawn from two newspapers, two academic discussions, two novels and a set of minimally-structured telephone conversations.

Answer 4

4. When and How should I use the perfect?

SHORT ANSWER:

☛ Use perfect constructions to introduce prior eventualities as context for the current discussion.

LONG ANSWER:
In the survey of perfect meaning we encountered several rules about when and how not to use the perfect (not with the HAVE piece in the progressive or perfect, not as narrative, not as the complement of a Wh-cleft, and so forth), but little or nothing about when the perfect should be used. And this appears to be what gives learners the most trouble. Probably half of our questions about the perfect have to do with the choice of constructions—particularly the choice between present perfect on the one hand and simple past or simple present on the other.

Unhappily, there is no hard-and-fast rule. The choice of a perfect to introduce a prior eventuality, like the interpretation of a perfect, depends on context—not just the nature of the eventuality introduced, but the reason it is introduced and the temporal context into which it is introduced.

So the fact that a perfect is used in particular circumstances cannot be generalized to a rule that it should be used in those circumstances. For instance, many learners are under the impression that because a past perfect is often used to speak of one event being prior to another, it should be used whenever you do so. Now it is true that there are circumstances when you must employ the past perfect:

^OKAt the time his first play was produced, Shaw had already established a substantial literary reputation.

But what ‘requires’ the past perfect there and forbids a simple past is not the time sequence but the adverbials at the time his first play was produced and already, which both locate the later endpoint of the Event Time timeframe at Reference Time. Without those adverbials it is quite possible to express the same time sequence in a sentence which permits either a simple past or a past perfect:

^OKShaw had established a substantial literary reputation before his first play was produced.
OR
^OKShaw established a substantial literary reputation before his first play was produced.

In terms of literal meaning, these two sentences amount to the same thing.

So the content of the sentence is not a reliable guide to whether or not it wants to be expressed as a perfect construction or something else. Take a look at some of the ELL questions and answers in §5, particularly those under the heading Questions about choosing between perfect constructions..., and you will find that each case is different, and that answers to each question are often very diverse, and even contradictory—because the context is inadequately specified.

This is why I am so fond of an answer here on ELL which I have christened FumbleFingers‘ Perfect Truism. (FumbleFingers speaks specifically of the past perfect, but the principle may be generalized).

☛ ”Don’t use the perfect unless you need it.”

What governs the use of the perfect is not the content, the meaning it expresses, but the purpose it serves. If you want to know whether to use a perfect, look at what you are trying to accomplish. What really distinguishes the perfect from the deictic constructions is focus: are you talking about state of affairs current at Reference Time or are you talking about the prior eventuality which in some sense gave rise to the current state of affairs?

An analogy from the cinema may help you understand this.

4.2 A cinematic analogy

Most of the action in a movie is told with medium shots and closeups. The camera is framed narrowly on one or two actors, or on a significant detail, with the background often in soft focus, reduced to mere visual texture. Here are examples from a famous gangster movie, Little Caesar (1931).

But this sequence starts with a wide shot: there are several actors in the frame, and the background, the setting, is in sharp focus.

This is a classic establishing shot: it establishes the spatial relationships between the actors, and between the actors and the setting. It visually defines the following scene as a confrontation between the rising hoodlum Rico and the established boss Diamond Pete.

A different use of the wide shot appears in the opening scene of another famous gangster movie, The Godfather:

In this case, the wide shot is a five-second cutaway which occurs more than three minutes into the scene. It suddenly reveals that what we have taken to be an intimate two-scene, witnessed only by a faceless functionary, is in fact a performance “starring” Don Vito and stagemanaged by him to humiliate Bonasera before the Godfather’s courtiers.

4.3 The perfect as wide shot

Perfect constructions work the same way. Think of the verb constructions in a narrative or discussion as a sort of ‘time camera’: just as the movie camera shows events and relationships in space, the verbal construction shows events and relationships in time. In this metaphor, Speech Time is where the camera stands ... Reference Time is the direction where the camera points—past, present or future ... and Event Time is where the ‘action’ takes place.

• Deictic constructions—simple and progressive —act like medium or closeup shots: they present consecutive single eventualities, without regard to their background.

• A perfect construction, on the other hand, is like a wide shot: it defines temporal relationships, presenting a prior eventuality not in itself, as an event, but as a state, the background or context for what is currently narrated or discussed.

You don’t want to push the analogy too far—a wide shot can include narrative action, as a perfect cannot. But the purpose of both a wide shot and a perfect is to shift your perspective and show you the larger scene—spatial in the cinema, temporal in a text—within which the action takes place.

Like the wide shot, the perfect serves both ‘cutaway’ and ‘establishing’ uses. Here, for instance, is a cutaway use in the opening sentence of a scholarly biography:

The room on the first floor of the Barbour County Courthouse in the little town of Eufaula, Alabama, was normally the County Clerk’s Office, but after it had closed for the day on August 2, 1957, it was being used by the county’s Board of Registrars, the body that registered citizens so they could vote in elections—not that the Board was going to register any of the three persons who were applying that day, for the skin of these applicants was black.

The closing itself is not important, and the author is not concerned to tell us when or how or by whom that happened. This passage is about the time after closing; the author mentions the closing in order to evoke a contrast between the ordinary use of the room for everyday public activity, and its after-hours use by a much smaller group for other purposes—purposes nominally ‘public’ but in fact a private conspiracy to subvert ordinary public decorum.

In non-narrative discourse the cutaway perfect is typically employed to introduce prior eventualities as evidence for a current position or as the antecedents of a current state:

Previously, scientists assumed that every occurrence was causally explainable, but now quantum physics has shown that this assumption is not true. Some kinds of occurrences are random and uncaused.

The European Union has tended to enlarge along regional lines, adding groups of nearby nations.

• Note that in the sciences, the cutaway perfect is preferred in citations—“Nishiyama and Koenig (2011) have observed that”—but in the humanities a present is more usual—“Nishiyama and Koenig (2011) observe that”.

An establishing perfect is used to introduce prior events as a current topic:

Many L2 researchers have criticized Krashen’s theory. One of the most devastating and detailed critiques came from McLaughlin (1987), who pointed out that ...

There have been three major reactor accidents in the history of civil nuclear power—Three Mile Island, Chernobyl and Fukushima. One was contained without harm to anyone, the next involved an intense fire without provision for containment, and the third severely tested the containment, allowing some release of radioactivity. These are the only major accidents to have occurred in over 14,500 cumulative reactor-years of commercial nuclear power operation in 33 countries.

Note the tense-shifting which such uses permit. The perfect construction anchors the discourse in the present and then leaves the author free to move his discourse between past and present tenses. This is the characteristic pattern with “Hot News” perfects:

   Calgary Police have laid a charge of first degree murder against a Calgary man in connection with the death of Graham Sear last June.
   Christopher Brian Richards, 30, was arrested on Wednesday in Beacon Hill.
   Graham Sear was found dead inside a silver Toyota 4Runner SUV on June 8, 2012, that was parked behind the Braeside Shopping Centre.
   Witnesses say they saw an assault inside the vehicle and spotted two men in their 20s fleeing the scene shortly after.

Nishiyama and Koenig point out a related ‘Topic Negotiation’ pattern in an example from a corpus of telephone conversations. The participants, strangers to each other, labor to find something to talk about, and each possible topic is introduced as a perfect:

A: Have you seen Dancing With Wolves?
B: Yeah. I’ve seen that, , that’s, uh, that was a really good movie. Probably one of the best things about it was the scenery and, uh, I thought the story was pretty good, too. I, I think Kevin Costner did a really good job with it.
A: Have you ever lived in that part of the country?
B: No. I haven’t.
A: Have you ever visited it?
B: Um, I’ve visited the Wyoming area. I’m not sure exactly where Dances With Wolves was filmed.
A: I think it was the Black Hills of South Dakota.
B: Could be. I, n-, I haven’t been to South Dakota. Have, have you been up to that?
A: Well, I lived in Omaha ...

Time-shifting is not restricted to present perfects; it may also be employed with past and even future perfects:

By the end of October, 2013, 95% of my students will have completed their first project and first written reflection, following a rubric. Evidence of their progress will be documented in their Developmental Workbooks (DW). Feedback throughout the project will also be evident in their DW's. By December, they will have completed their second project.

In all of these cases what is going on is that a perfect is employed to introduce prior eventualities while maintaining an ‘anchor’ in the current Reference Time. The perfect is a signal to the reader or hearer that the prior eventuality is being introduced not for its own sake but as background or context for the current topic. With a cutaway perfect, the speaker returns to that topic immediately, while with an establishing perfect the speaker may digress into discussion of the prior eventuality for several sentences or even pages. But the initial perfect acts as a sort of ‘back button’: it maintains the possibility of returning to the original timeframe and the original topic, equipped now with information about the context which provides a deeper understanding. As Nishiyama and Koenig put it (supporting the assertion with a complex formal description which is far beyond the scope of this essay):

Why do writers or speakers choose a perfect form to describe an eventuality that occurred or started in the past? Simply put, our answer is that the choice of a perfect form is guided by writers’ or speakers’ desire to help addressees understand the coherence of the discourse they read or hear.

I suggest then that you use the perfect as a structural device:

☛ Use perfect constructions to introduce prior eventualities as context for the current discussion.

In other words:

☛ ”Don’t use the perfect unless you need it.”

This is the Golden Rule. Everything else is commentary.

NEXT: 5. ELL questions about the perfect

Answer 5

5. ELL questions about the perfect

This is not a complete list of ELL questions addressing the P_^F but it should help you find questions similar to yours.

Questions about constructing a perfect correctly
   Haven't eaten/haven't ate
   Confusion with had + forget
   "I have ever went to"
   "Is not being" vs "Has not being"
   "Had remained" vs "Had been remained"
   Using "verb + ed" for the future
   "I done" vs "I have done"

Questions about distinguishing perfect constructions
   "Read and accepted" or "read and accept"?
   People who have had their belongings taken.
   What role does this PaPpl take?
   Can a living person "have left" a family?

Questions about perfect tense
   Why is the past perfect used in this sentence?
   How often do you use FuPf?
   Application of 'will' + PrPf constructions
   Difference between "has + verbed" and "had + verbed"
   Time references with Pa and PaPf
   Can PrPf be used with a dead subject?
   Pf used to say one action happened before another

Questions about perfect aspect
   Does 'had been' imply ongoing action?
   Does Pf imply incompleteness?
   Does PrPfPrg imply temporary action?
   Is Pf appropriate with "forget to"?
   Is Pf appropriate with "break this rule"?
   Does "how do you know" have Pf implication?

Questions about modal, irrealis and conditional perfects
   How could he have been so stupid?
   I would not now have exchanged …
   I might/may have sat there
   I couldn't tell/have told if he was annoyed
   Might have been but wasn't
   My dog could jump/could have jumped
   "be going to" in FuPf
   He will have been drinking all day
   He had to have arrived
   I may beat him now/tomorrow .. I might beat him a year ago
   Subjunctive had/had had
   If I were … I would have …
   Modal conditional perfect
   Would have been/caused
   Conditionals with PrPf consequence
   Conditionals with modal Pf consequence
   If I were … I would have made
   If they had gone … I'd rather you had gone

Questions about the kinds of perfect meaning
   Something I may still do - haven't or didn't
   Simple past with consequences
   Because he told/has told me so.
   Can PrPf be used with a stopped action?
   PrPf or simple past?
   What tense should I use?
   Why is it not OK to use the PrPf here?

Questions about choosing between perfect constructions and simple present or past constructions
   I never met him/have never met him
   PrPf/PaPf/Pa with no relation to present
   Which tense with recent past?
   is/has been on the news
   I am/have been a fan
   is/has been a tough day
   issues we face/have faced
   Did not expect/Had not expected
   Is being/has been developed
   Why didn't he make/hasn't he made the decision

Questions about coordinating the use of perfect constructions with clauses using other verb constructions
   Before he came/had come .. Tom had studied/was studying
   Making sense of shifting tenses
   When is PaPf not necessary?
   What tense(s) with successive past events?
   May we mix different tenses?
   When is the PaPf exactly needed?
   Chronological order of tenses
   I met/had met you before you met/had met her
   Tense in a clause embedded in a PaPf main clause
   Should I use the PaPpl of every verb that follows 'would have'?
   This is/was/will be the first time I have been/had been/go to
   PaPf with past events out of order

Questions about the use of perfect constructions with time expressions
   PaPfPrg with 'during'
   Pf with ‘as soon’, ‘when’
   Pf with 'ever'
   PrPf with 'until'
   PrPf with 'ever since'
   Tense with 'so far'
   Does 'before' make the Pf redundant?
   PrPf with past time
   Pr with 'always'
   Was/has been injured 'for a week now'
   Pf with 'for' but now ended

Answer 6

1. How do I construct a Perfect?

SHORT ANSWER:
☛ English perfects have two components: a form of the auxiliary verb HAVE on the left (the ‘HAVE’ piece), and a verb ‘group’ headed by a past participle (P_^A⋅P_^PL) on the right (the ‘VERB’) piece).

---

LONG ANSWER:
Detailed rules for constructing perfects, with examples, may be found here. Briefly:

• English ‘perfects’ employ a form of the auxiliary verb HAVE followed by the past participle (P_^A⋅P_^PL) of the next verb. In a ‘simple’ P_^F construction that next verb is the lexical verb (VERB), the verb which carries the ‘meaning’. Here, for example, is how that works with the lexical verb WRITE:

  

  A typical ‘simple perfect’ sentence may be diagrammed like this:

• More complex perfect constructions combine the perfect with modal verbs (MODAL) and with progressive (P_^RG) and passive (P_^SV) constructions. The rules are basically very simple:

  • These constructions follow a strict order: the MODAL component first, the perfect component next, then the progressive, then the passive, with VERB always at the end.
  • Each construction is marked with a specific auxiliary verb, HAVE or BE, and there is a ‘ripple’ effect: the form (present or past participle or infinitive) of each of these verbs is determined by the preceding component.

Here’s the diagram for a modal perfect progressive version of the sentence diagrammed before:

1.1 Common construction errors

One common error in building these constructions is to use the simple past form of VERB instead of the past participle.

QUESTION 4026: (user hjpotter92)
While chatting with somebody, I first used haven’t eaten, then I thought that it was wrong, and switched to haven’t ate.

Apparently, haven’t ate is the one which is wrong.

Can someone explain the logic behind this? Verb tenses are still something I confuse a lot.

Eaten is the P_^A⋅P_^PL of EAT which should be used here; ate is the P_^A form.

Another common error is omitting the HAVE form:

QUESTION 9532: (user Jader Dias)
I think this sentence is correct.

• I done something.

This is also correct.

• I have done something.

What are the involved tense? How are they different?

One obvious cause of these errors is the fact that for regular English verbs the simple past and past participle forms are identical: the form with the –ed suffix. They are identical for many irregular verbs, too: felt, laid, meant. But for many very common verbs the two forms are different.

Any good dictionary will tell you the past participle of a verb if it is irregular. Wikipedia lists the principal parts of the most common English verbs here, and what claims to be a complete list here.

Another contributing factor is that native speakers often make the same mistake—or what sounds like a mistake:

• Sometimes it’s a matter of dialect. Local speech throughout the English-speaking world uses different non-standard verb forms.
• In some cases, standard use itself varies: in North America, for instance, the past participle of get is gotten, while elsewhere it is got. For some verbs, there are both regular and irregular forms in standard use—the language is always changing.
• Sometimes you hear a mistake which isn’t really there. In speech the HAVE form is unstressed, so in rapid speech it isn’t just contracted to -’s or -’d or -’ve , it may be actually slurred to the point that it can’t be heard at all.

But usually the mistake is just a mistake. Don’t imitate it in either speech or writing.

The complex constructions can be even more confusing for the learner:

QUESTION 7820: (user muhammad)

• BBC channel is not being broadcast in our area for a year ago.
• BBC channel has not being broadcasted in our area for a year ago.

Which one is grammatical? Depending upon time span I think number two is grammatically correct.

One reason is that BE is the auxiliary for both progressive and passive constructions, so it’s easy to get them mixed up.

Another reason is the ‘ripple’ effect: there isn’t one distinct form for each component in a construction. Instead, the form which a component takes depends on what comes before it.

Finally, there’s the fact that learners often try (or are compelled) to learn the complex constructions before they have completely mastered the underlying basic constructions. In the question above, for instance, it is clear that the questioner hasn’t really grasped the difference between

• the progressive construction, which employs the auxiliary BE with a following present participle
• the passive consruction, which employs the same auxiliary BE with a following past participle
• and the perfect construction, which employs the auxiliary HAVE with a following past participle

There’s really no silver bullet for this. You need to have a clear idea of what the different constructions mean, so you know which construction you want to use. And then you need to write … read … speak, over and over again, in contexts where your errors are identified and corrected. Eventually, after you’ve used a construction three or four hundred times, it will become second nature.

NEXT: 2. How do I recognize a perfect in context?