logo  ELT Concourse teacher training
Concourse 2

Reported or indirect speech


This first section mostly reiterates material in the initial training section and is here as a reminder of the basics.  You can skip this if you are already aware of the basic issues or have recently worked through the initial training section for this area.
If that is the case, skim through what follows, and/or do the mini-test and then move on.  It's up to you.

chess player

Consider these pairs

I was in London He said he had been in London
I can help She said she could help

On the left we have the direct speech – the words uttered.
On the right we have reported or indirect speech – how the message is passed on.

On the face of it, there's nothing terribly difficult about this idea.  The tense shifts back one (from, e.g., was to had been, from can to could)At the same time, I changes to he, we changes to they and so on.
Here's a list of the changes in English.

Language item Change
Present simple changes to past simple I am in London
He said he was in London
Past simple, present perfect and past perfect all come out as past perfect I was in France
She said she had been in France
I have been to France
She said she had been to France
I had been to France
She said she had been to France
Present progressive changes to past progressive I am writing to him
He said he was writing to them.
Future 'will' changes to 'would' I'll go later
He said he'd go later
Future 'will be'+ ing changes to 'would be' + -ing I'll be seeing him tomorrow
She said she'd be seeing him tomorrow
Future 'will have' + past participle changes to 'would have' + past participle I'll have done it
He said he would have done it
Other changes
Pronouns change as appropriate I / We / want to be there
I / He / She / We / They said I / she / he / we / they wanted to be there
Time and place expressions change as appropriate I like it here
She said she liked it there
I am going tomorrow
He said he was going the next day
Modal verbs change to their 'past' equivalents if there is one I may see him
He said he might see him
I must go now
He said he had to go then


Using common sense

Of course, not all changes are always appropriate (but using the changes will usually be correct).
If we are reporting something virtually simultaneously, then we often don't change the tense or time expressions.  If we are reporting something in the same place, then we don't change the place expressions.

So we might get:
    A: I'm going there now.
    B: What did he say?
    C: He said he's going there now.

If an utterance remains true, we often don't change the tense so we get, e.g.,
    I'm from South Africa = He said he's from South Africa
    I love the countryside =
She said she loves the countryside

Try this matching exercise to make sure you have understood so far.

Did you notice the changes, particularly with time and place expressions but also with the verb come (which changed to go)?

If you have followed so far, this will be familiar:

1 "I was in London," he said 4 He said he had been in London
2 "I can help," she said 5 She said she could help
3 "You are welcome to come.  Would you like to bring Mary?" he said 6 He said I was welcome to come and would I like to bring Mary

It's clear that we have examples of direct speech and indirect speech here in sentences 1, 4, 2 and 5 but Sentence 6 is what is called a hybrid form because the first part follows the 'rules' but the second part actually changes only the pronoun, from you to I.
If the sentence followed the reported speech 'rules', it should be
    He said I was welcome to come and asked if I would like to bring Mary
which is another possibility, of course, but sounds quite formal.


Reporting closed questions with if and whether

"Were you there on the night?" They asked whether I was / had been there on the night

Closed questions are those which require a Yes or No response and they are usually reported with if or whether.  We get, therefore, for example:
    Are you going to the cinema?
reported as
    He asked her if she was going to the cinema

There is a bit more to it, however.

Consider what direct speech is being reported in the following.

  1. I asked whether there were any good recipes in it.
  2. I asked if there were any good recipes in it.
  3. She wondered whether to go.
  4. He asked whether or not they could come.

When you have done that, try reporting these sentences (from the point of view of later and elsewhere).
Then click to reveal the comment.

  1. “Are they English?”
  2. “I wonder if they are English.”
  3. “I’m wondering whether to join you.”
  4. "Can I talk to you?"


Reporting open wh-questions

Who are you looking at? It asked me who I was looking at

Open questions are usually phrased using wh-words: who, what, why, where, how.  They cannot be predicted to have a Yes-No-Maybe answer.
Questions formed in this way cannot be reported with if or whether.  The reporting is done by embedding the questions.

Embedding is often associated with polite questioning so, instead of the direct:
    Where is the station?
we form polite embedded questions such as
    Can you tell me where the station is?

So it is with reported questions.  Thus:

Where is the station? He asked me where the station is / was
What is the time? She asked me what the time was / is
When are you arriving? He asked me when I was / am arriving
Who is coming with you? She asked them who was / is coming with them
Why are they late? She asked why they were / are late

The tense chosen will often conform to the time and place of the reporting using the common-sense rules discussed above although back-shifting tenses where possible is common even when the reporting is virtually simultaneous.

The big issue for learners with this kind of reporting is the ordering of the subject and verb.  Most language will lead learners to produce errors such as:
    *Can you tell me when is the film beginning?
    *Do you know who is the lady there?

And this will also carry over to reported questions so we get:
    *She asked me where is the zoo
    *They enquired what time did the train leave

and so on.

Other reporting verbs such as explain, clarify, complain, mention, remember and state will produce similar errors because the structures are parallel to reported questions in English but not paralleled in many other languages.  We may encounter, therefore:
    *She explained how did the machine work
    *They clarified what did they need
    *I remembered where was I going

See below for more on reporting verbs.


that and what

What are the rules for using that and what in reported speech?  Report the following using that or what if possible and then reveal the commentary.

  1. "I am coming now."
  2. "I don't know her name."
  3. "What's your name?"
  4. "My name is Mary."
  5. "I will not go if it rains"


Tense shifting

As we saw in part 1 of this guide, tense shifting is common in English and it is rarely wrong to do it.  However:

  1. If the reporting verb is in the present, we don't shift tenses.  So we get
        She often says, "I don't know what to do"
    changing to
        She often says she doesn't know what to do
        "There was a nasty accident here last night," John informs me
    changing to
        John tells me there was a nasty accident here / there last night.
    (Note that last night does not change because the reporting is of a recent utterance.)
  2. If the validity of what was said still holds.  For example:
        Darwin wrote, "There is grandeur in this view of life."
    changes to
        Darwin wrote that there is grandeur in this view of life.
    not to
        Darwin wrote that there was grandeur in this view of life
  3. Although back-shifting could be used in all the above examples, there are rare times when it actually produces nonsense.  Try reporting
        "I chose to study French because it was a beautiful language."
    Will you accept
        He said he had chosen to study French because it had been a beautiful language?


Reporting commands and exclamations

We saw above that question forms are reported differently from statements.  How would you report these?  Click here to reveal some comments.

  1. "What awful weather!" she exclaimed.
  2. "Stop fidgeting!" she said to John.
  3. "Stop fidgeting," she growled.



Here's a definition:

The name given to those aspects of language whose interpretation is relative to the occasion of utterance
Fillmore (1966) in Harman (1989)

It's an important phenomenon in this area because the use of deixis neatly explains a lot of the so-called anomalies of indirect speech.  Because meaning is dependent on the point of view, time, location and place of the speaker / writer we are obliged (or not) to change, e.g., bring to come, come to go (and go to come), this to that, here to there, yesterday to the previous day, now to then, bring to take and so on.
In this regard, the following changes now make more sense:

"I want you to come to me now." He said he wanted me to go to him then
"I'll go next year," she said She said she would go the following / next year
"I want to bring this food with me," he said He said he wanted to take that food with him

Once again, we find that context makes meaning.  The common-sense rule (see above) is now explained.

For more, there is a guide to deixis on this site which includes a larger image of the wheel above and explains what it all means.


Reporting verbs

Essentially, there are three types.  Can you categorise this list into three groups?  Click to reveal, as usual.

say cry ask
tell order offer
invite shout explain


Modal verbs

To sum up:

  1. If a modal has a past equivalent, it may backshift to that so we get may (for permission) / might and can / could, have to / had to.  Thus:
    You may go She said we might go (then) or She said we may go (now)
    I can play the piano She said she could play the piano (then) or She said she can play the piano (timeless enduring ability)
    I have to see the doctor She said she had to see the doctor (then) or She said she has to see the doctor (now / future)

  2. Those without past equivalents (could, might, ought to, used to, should) remain unchanged.  Thus:
    I could help He said he could help (then or now)
    That might help He said that might help (then or now)
    I ought to see the doctor He said he ought to see the doctor (then or now)
    I used to take my holidays there He said he used to take his holidays there (a discontinued past habit)
    You should work harder He said I should work harder (current mild obligation)

  3. The modal may / might is slightly confusing because it is used for both permission (when may is often back-shifted to might) and for possibility when no change is apparent.  Thus
    You may go She said we might go (permission given in the past)
    I may talk to her She said she may talk to her (concerning a current future possibility) or She said she might talk to her (concerning a possibility in the past)
    I might go for a walk She said she might go for a walk (concerning a slightly less likely possibility then or now)


Teaching issues

There is nothing very difficult about the form of reported speech changes (providing a learner is already familiar with the tense forms of English).  However:

  1. Because of the 'common sense' issues touched on above, you need to make sure that the language is very clearly set in a time-and-place context.
  2. It is almost impossible to practise the form changes in class by getting students to report each other's utterances because time and place remain static.  You need to spread the practice over time and place to be authentic.
  3. You need to make sure that learners are aware of the common-sense issues and don't slavishly transform every utterance.
  4. Languages deal with the issues differently.  Some, for example, reserve a subjunctive tense for reported speech and some hardly make any changes at all.


Teaching ideas

Teaching the mechanics of indirect speech is not too challenging providing the learners have a grasp of the tense forms and pronoun systems but one does need to address different forms separately or it all becomes a mass of data that bewilders learners.
A sensible approach is to apply the analysis as above, focusing on reporting declarative statements, yes-no questions, open, wh-questions, exclamations, orders and so on separately before making any attempt to combine ideas.

Here's an idea for teaching indirect speech and still applying the common-sense rules.

Step 1:

Use a little task sheet like this:
Write three things which are true about you now: Your name: ________________
Always true Example: I like reading Your sentence: ____________________________________
The past Example: I studied Italian at school Your sentence: ____________________________________
Tomorrow Example: I am having lunch at home tomorrow Your sentence: ____________________________________
Now collect all the papers and store them away for at least two days.
Step 2:
Distribute the papers randomly but making sure the learners don't get their own.
Now we come to the crux:
Statements like
    I enjoy classical music
can be reported in two ways (both correct):
    Sheena wrote that she enjoyed classical music
    Sheena wrote that she enjoys classical music
Other statements will need to be amended according to the rules of deixis in English so we might get:
    Raoul wrote that he was going to the cinema the next day
    Michel wrote that he had eaten macaroni for lunch that day

Obviously, this is contrived and artificial to some extent but it is personalised and situates the language temporally and spatially.  It is certainly better than meaningless sentence-transformation exercises.
what was the question

The word-order issue: what was the question?

Because the word order when reporting questions and using a number of the reporting verbs is a common source of error, it is worth practising separately.  Fortunately, the use of back-shifting, even for virtually simultaneous reporting, is also common so there is less need to set up delayed reporting.

Step 1:
Focus on an area in which it will be natural to ask lots of questions.  Focusing on the teacher's life experiences is a good way to begin so that might evince questions like
    Where did you go to school
    Why did you decide to be a teacher?
    Where have you taught?
    Do you enjoy teaching?

Depending on the level of the learners a list of 20 or so things they want to know about the teacher is not too ambitious.
Step 2:
Once the learners, in pairs or individually, have their lists of things they want to know and have properly formed questions they can ask them in open class.  Unfortunately, at this point, the teacher becomes somewhat deaf and is forced to ask another student what the question was so we get, for example:
    Petra: Where did you train to be a teacher?
    Teacher: I didn't get that.  What was the question, Maria?
    Maria: She asked you where you trained.
and so on.
Step 3:
Now, in groups of three, the learners play the same game, deciding on what things they want to know about the others in their group and having one questioner, one slightly deaf responder and one reporter in each trio.  They all get a turn at playing each role.
Note that:

  1. You have to teach the forms before you can launch into this kind of practice and
  2. The questions may well be mixes of wh-questions and closed questions so the reporting will include formulations such as
        He asked you why you became a teacher
        She asked you if / whether you enjoy teaching

You can set up the task to exclude one or other type of question, of course, but that's a lot less natural.

Related guides
nominalised clauses for an alterative way to look at reported or embedded questions inter alia
a list of reporting verbs for a list of the commonest possibilities
reporting verbs in academic writing for a guide to reporting what people said or wrote in EAP contexts
modality for the index to guides to this area
deixis for more on a key area

Click here for the test.

Chalker, S, 1987, Current English Grammar, London: Macmillan
Harman, I P, 1989, Teaching indirect speech: deixis points the way, English Language Teaching Journal, Volume 44, No 3, pp230-238, Oxford: Oxford University Press