Reported speech in English is a complicated matter, and then again
it’s quite simple.
It is complicated when we try to interpret sentence pairs which have
been made up (by the grammarian or non-grammarian). It is simple (well,
to a point) when we WATCH the reporting of speech in its natural habitat,
i.e., when we take into account that reported speech is used in communication—just
like any other language item. That is what language is for, after all.
Communication takes place in communication situations.
A communication situation has two main parts: (1) the speech situation,
which contains speaker, hearer and the language forms used; and (2)
the reference situation, i.e. the situation which is being discussed,
it contains the referents, which the language forms refer to or point
to ("house" ---> building). Both, speech situation and
reference situation are located in space and in time, in real life or
in fictional life, and so on. What is being said and how it is being
said in a communication situation is influenced by speaker, hearer,
referents as well as by its locational parameters.
Let us look now at the communication situation which is connected with
the phenomenon of reported speech. The difference between a "regular"
communication situation and one that has to do with reported speech
is that what speaker and hearer refer to is itself a communication situation.
Now when speaker A and hearer A discuss the embedded communication
situation B, there are lots of things that can be referred to: speaker
B (e.g., estate agent), hearer B (owner of the house), language form
B ("house"), the communication activity (shouting at each
other), referent B (dilapidated building), where communication situation
B takes place (in Sydney), when it takes place (three years ago), and
so on. The language form "house" is just one of many things
that can be discussed. So the term "reported speech" is utterly
misleading (and so is "indirect speech," of course), it should
be something like REPORTED COMMUNICATION (SITUATION).
Basically, then, there is nothing special about reported communication.
As always: the words we use and the grammatical structures we use in
situation A (including the tenses and the pronouns and adverbs) depend
on who is communicating when and where and what they want to say.
Coming to "Paul went to the movies yesterday": This is a
sentence which has been artificially made up, it is as dead as a doornail,
it has never lived, it has no speaker, there is no addressee, there
has never been a communication situation "around" it.
To "report" it is simply impossible, because we can’t
discuss a non-existing communication situation. Period.
However, that’s what we do—what grammarians have been doing
for I don’t know how long. How do we go about this impossible
First, we strip the language form (B) of everything it is connected
with, the communication situation it is part of, the referent it refers
to, the speaker who has produced it, etc., i.e., we reduce it to its
Second, we pretend that human memory is much better than it actually
is. Speaker A, who reports what speaker B has said, is supposed to
be able to remember the exact wording of language form B. Research
has shown that this is not so. My own experiments with students who
had to report situations immediately after they had happened, a week
later and again two weeks later, showed that even immediate reports
rarely used the same words as the original speaker B. They just didn’t
remember. This applies also to the original speaker himself.
Next, we pretend that we report speaker B’s speech. What we
do in real life, however, is that we discuss the whole communication
situation B and its elements (see above). Where we refer to reference
situation B, we use the words WE want to use, which are not necessarily
those that speaker B has used.
Finally, having only (the form of) a sentence left, we invent purely
formal transformation rules like backshift (one tense back) and: "yesterday"
must be changed to "the day before," "here" becomes
"there," first person will pop up as third person, and what
And now that reported speech is stripped of any link with real life,
we are ready to freely juggle with sentences, whether they have been
killed or have been dead right from the beginning.
Thus our stillborn "Paul went to the movies yesterday" will
be transformed into:
"I said Paul had gone to the movies the day before,"
"I said Paul went to the movies yesterday,"
"I said Paul went to the movies the day before,"
||"I said Paul had gone to the movies yesterday,"
Result (a) is the only one that the average school teacher in Germany
would accept— it’s the only one that conforms to the "rules."
But does it make sense?
Result (b) seems to make more sense, but violates the "rules."
Results (c) and (d) violate some of the "rules" and also
Now there are two types of grammarians, the rule-governed one and the
free-thinker. Even the free-thinker is not quite free in his/her thinking.
We all are contaminated by the rules we had to learn when we were young,
and we all are prepared to ignore the rules once in a while, the free-thinker
more so, the rule-abiding citizen less so.
Which means to say we try to save a sentence, even if it violates a
rule or two or doesn’t make sense–as long as there is the
slightest chance for it.
We do this
by bending the rules a little, e.g. by creating exceptions to
a rule (no backshift with eternal truths),
||by re-discovering communication, i.e. we take our dead sentence
pairs and try to build possible communication situations around
Either way works only so far.
Let me give two illustrations.
(a) The case of the missing backshift
Advertisement in the (British) Observer (7.3.74):
I TOLD my solicitor I WANT to help old people—and he showed
me how to multiply my help twenty-fold ... You SAID your help IS BEING
Len Deighton, Funeral in Berlin (1964):
I TOLD you I’VE BEEN PLANNING this for fifteen years.
Herald Tribune (2.11.72):
Senator Mansfield SAID that he INTENDS to keep his post as Senate
None of these items refer to eternal truths, and still the writers
do not use backshift. More exceptions will have to be described by the
grammarian, and so many grammarians do, until there are more exceptions
than rules. Which is rather impractical.
(b) Rachel’s "rescue" operation
Rachel, in order to "save" the two transformations sent in
by Orlando Carranza, creates the following communication situation:
SPEECH SITUATION A
speaker A: unknown
hearer A: non-existent
language form A:
I said Paul went to the movies yesterday
place of speech situation A: unknown
I said Paul had gone to the movies yesterday
time: 11:00 a.m.
REFERENCE SITUATION A
Communication situation A
Speech situation B
speaker B: (same as speaker A) unknown
hearer B: non-existent
language form B: Paul went to the movies yesterday
place of speech situation B: unknown
Reference situation B
Action: going to the cinema to watch a film
time: one day before speech situation B ("yesterday")
Most elements of the communication situation are unknown or missing,
in fact, Rachel has only invented the time of speech situation A (11
o’clock). The time of reference situation B is concluded from
what the unknown speaker A is supposed to be saying (yesterday).
What are the results of the rescue operation?
(a) I said Paul went to the movies yesterday.
Speaker A says at 11 o’clock that she said before 11 o’clock,
but on the same day, "Paul went to the movies yesterday."
This makes good sense, although, of course, we have no idea why on earth
the speaker would report on her own communication situation in the first
(b) I said Paul had gone to the movies yesterday.
Now this wording does not make any sense to me. There is nothing in
reference situation B that gives us any hint that something else happened
after Paul left for the cinema.
Rachel tries to solve this problem by introducing the idea of formal
and informal language. I myself don’t think that the sentence
can be rescued by calling it formal. I rather think that Rachel subjects
herself here to the rule-abiding one of the two souls that dwell in
her breast (goodness).
The trick of inventing a communication situation around dead sentences
is complicated and rather confusing, because the communication situations
which we are able to imagine are no more than fleshless skeletons with
lots of missing bones, which can never come alive, nor will they make
made-up plastic sentences come alive.
If we really want to understand reported communication, we should deal
with living communication situations and living sentences. If we do
so, i.e., if we study reported "speech" where it is actually
used, we will see that the FORMAL rules we were taught are utter nonsense
and can even lead to misunderstandings. Not that there are no rules,
there are rules galore for the rule-greedy, but they are universal rules
that apply to all communication situations not just to those that contain
an embedded communication situation, i.e. what we call reported "speech."
Back to Reported (indirect) speech (with past tense)