Facts Embedded In Mental Clauses

Halliday & Matthiessen (2014: 542):

There is no mental process corresponding to fact or chance, no implication of a conscious participant that is doing the projecting. Unlike nominal groups with nouns of projection, nominal groups with fact nouns are not nominalisations of projection nexuses. A fact, as already pointed out, is an impersonal projection.

However, it is possible for a fact to enter into a ‘mental’ process clause without being projected by it. In this case it functions as a Phenomenon within the mental process clause. For example:

The fact [[that he rides in such exalted company]] will not deter Scott. 

With the heavy expenditure on new rating, plus a new street costing 1,000,000, the cost of the Pump Room, new Municipal Offices, and so on, the eventual rates are likely to deter people from coming to live in the town, as they would probably be influenced more by excessively high rates than by the fact [[that there was a luxury swimming bath for use in winter]]. 

He overlooked the fact [[that Ceylon had to be governed not only in the first few years after independence but for all time]]; and this raises several questions. 

Sternberg himself photographed the film, revelling in such pure artificiality, regretting only [[that he had to use real water]]. 

You know I smoke and I hate it. I hate [[that I do it]]. And I’m at that point where I have to make the decision. I can’t go on any longer with it.

Facts Embedded In Existential Clauses

Halliday & Matthiessen (2014: 542):

Another, minor but significant, environment in which fact clauses occur is that of ‘existential’ clauses (an environment favoured by evidence):

There is evidence [[that the Russians were just as surprised as anyone else at the suddenness and violence of them]], but it is, of course, a situation ideal for exploitation. 

If the serum of a D negative individual agglutinates the D positive but not the D negative control cells, there is a high probability [[that the serum contains anti-D]], but the specificity should be confirmed by testing against several more examples of D-positive and D-negative red cells.

Facts Embedded In Personal Attributive Clauses

Halliday & Matthiessen (2014: 542):

However, we also need to take note of ‘attributive’ clauses where the Carrier is realised by a nominal group denoting a person and the Attribute is a nominal group with an embedded fact clause, either ‘possessive’ with a noun as Head (e.g. idea, notion, inkling [[that ...]]) or ‘intensive’ with an adjective as Head of the nominal group (e.g. sure, certain, aware, cognisant, oblivious (of the fact) [[that ...]]); for example:

They would have no idea [[[that the current British theatrical renaissance is having an effect far beyond the West End of London, || so that Broadway is heavily influenced by the highly successful plays of today [[that it has imported from Britain]] ]]]. 

However, I am not sure [[that [[what probabilists and what physicists mean here by ‘fields’ ]] are quite synonymous]].

These ‘personal’ ‘attributive’ clauses are closely agnate with projecting ‘mental’ clauses: they have no idea ~ they don’t know, I’m not sure ~ I don’t know.

Fact Clause Serving As Token In An Identifying Clause

Halliday & Matthiessen (2014: 541-2):

In an ‘identifying’ clause, the fact clause serving as Token is identified with a Value realised by a nominal group with a noun as Head that typically belongs to the class of fact nouns; this fact noun may itself be qualified by an embedded fact clause. The Value is an interpretation of the fact clause, identifying it as a particular fact of some class of fact such as reason, problem, lesson, difficulty. … the Value nominal group may include an Epithet (thorniest, most important, plain … ; cf. obvious, indisputable, appalling, significant, simple, mere) assessing the fact represented by the Token (in the same way as Epithets within Attributes do) or a Numerative (cf. first, next, last). This latter is important in the development of discourse, being agnate with an internal temporal conjunction (cf. thirdly, the supreme interest for the whole world ...): the enumerated Value is the thematic point of departure of the clause and this Theme locates the clause as a message in the unfolding text.

Attributes Ascribed To A Fact Clause Serving As Carrier

Halliday & Matthiessen (2014: 540-1):

In an ‘attributive’ clause, the Attribute ascribed to the fact clause serving as Carrier is realised by a nominal group with an adjective or noun as Head belonging to one of a small number of classes. These are illustrated in Table 7-27. Several of the types are similar to classes of interpersonal Adjunct and two of them can also be related to types of sensing in ‘mental’ clauses.³⁴ We shall return to this area in Chapter 10, Section 10.3.1, showing that certain ‘attributive’ clauses with fact clauses as Carrier and Attributes of assessment form part of the realisation of a semantic system of assessment. The nouns in it is include fact nouns such as fact, idea, but they also include nouns of evaluation such as pity, shame, nuisance that are less likely to function as Head/Thing in nominal groups with a fact clause as Postmodifier/Qualifier.

∞

³⁴ The ‘attributive’ clause may have an agnate ‘mental’ clause of the “please” type: it is surprising (to me) that ... ~ it surprises me that ...; the equivalent of the Senser in the ‘mental’ clause is a circumstance of Angle

The Typical Environment For A Fact

Halliday & Matthiessen (2014: 540):

Other than with impersonals such as it is said, it is rumoured, it seems, the typical environment for a fact is a ‘relational’ process clause of the ‘intensive’ type, either ‘attributive’ or ‘identifying’, e.g.

[attributive] 

Earl Russell says it is inevitable, though profoundly regrettable, [[that the agitation against the Polaris base has generated some antagonism to the policy of the United States]].

In that article, it’s no coincidence [[that I have a big fight with Twain and Eliot]], || because I disagree with them on issues [[that concern all of us]].

Until 1940 it was an observable fact [[that there were composers whose music was highly prized in some countries and entirely neglected by their neighbours]], and this was explained by the difference in national characters.

It is clear [[that the Princess and her husband are settling down in London]] and for this purpose, Kensington Palace is well suited.

[attributed variant] 

The Federal Government has made it clear [[that it would have no part in any project for the development of long-range missiles – which in any case would contravene the provisions of the Brussels treaty]].

[identifying] 

The third reason is [[that the supreme interest for the whole world – East and West and uncommitted nations – is the prevention of nuclear war]].

The lesson [[that’s learned]] is [[that they aren’t Kangan]]; Kangan is everybody, as represented by the people gathered in Beatrice’s apartment at the end of the novel. [Text 16]Perhaps the most important point of all is the fact [[that capital was available for expansion as required]].

The plain fact is [[[that it is extremely difficult [[for MPs to accept invitations from foreign Governments, or from public relations organisations [[working for them]] ]], || without being compromised]]].

[identifying clause of proving] 

But the fact [[that they are caught]] proves [[that they do not lift above the headline]].

Here the fact is an embedded clause standing as a nominalisation on its own, functioning as the realisation of an element in the relational process clause (Carrier or Identifier/Token, in these examples).³³ Since it is embedded, there is always an agnate version where the fact clause serves as a Qualifier of a noun of the ‘fact’ class, e.g. the fact that Caesar was ambitious.

∞

³³ Strictly speaking the embedded ‘fact’ clause functions as Head of a nominal group which, in turn, functions as an element in the ranking clause. This analysis shows how clauses serving as Head are agnate with clauses serving as Postmodifier in nominal groups with a fact noun as Head: that Caesar was ambitious is obvious: the fact that Caesar was ambitious is obvious. But since a fact clause functioning as Head takes up the whole of that nominal group we can just as well leave out that stage in the structural analysis and show it as directly embedded into the clause.

Impersonal Projections Of Facts

Halliday & Matthiessen (2014: 539):

While there is no participant doing the projecting – no Sayer or Senser – a fact may be projected impersonally, either by a relational process (‘it is the case that ...’) or by an impersonal mental or verbal process, as in

[i] relational

it is/may be/is not (the case) that ...
it happens (to be the case) that ...
it has been shown/can be proved (to be the case) that ...
it happened/came about that ...

[ii] mental: impersonal

it seems/appears/is thought (to be the case) that ...

[iii] verbal: impersonal

it is said/rumoured (to be the case) that ...

Here the it is not a participant in the projecting process but is simply a Subject placeholder; hence the fact clause can occupy its position at the front: that Caesar was ambitious is certainly the case/is widely held/is generally believed, etc. By contrast we do not normally say that Caesar was ambitious was thought/said by Brutus – at least not in a reporting context, only in the special sense of ‘these lines were spoken by ...’; and this is because, as we have seen, where there is a personal projecting process, mental or verbal, the clause that is projected by it is not embedded but hypotactic.

Blogger Comments:

Lest this wording be misunderstood, the impersonal projection of a fact occurs within a single clause, since the fact is embedded; the projection is not a relation between ranking clauses in a clause complex.

A Fact Clause Serving As Head In A Nominal Group

Halliday & Matthiessen (2014: 538-9):

A fact clause serving as Head in a nominal group without a fact noun can be related to the first class of fact noun – that of ‘cases’, since such a fact clause is always agnate with an expanded nominal group with fact as Head. Whether the nominal group has a fact noun as Head or not, the fact clause is embedded. Because there is no projecting process involved, to which it could be paratactically or hypotactically related, a fact can appear only in embedded form: either as Qualifier to a ‘fact’ noun, or as a nominalisation on its own (Figure 7-23); for example:

Historically, the fact is [[[that Uncle Tom’s Cabin was the most popular novel of the nineteenth || century and had a huge effect on American history]]].

Fact Nouns And Reference

Halliday & Matthiessen (2014: 538):

Like nouns of projection, fact nouns can be used anaphorically (or cataphorically) to create cohesion in discourse; for example:

Warwick Town Council originally decided to build its own crematorium, but in April last year it abandoned the idea and entered into a joint scheme with Leamington Town Council and Warwick Rural District Council. 

The Bill is short and modest in scope, and it is doubtful whether the other Private Members’ Bills in the offing will fill all the gaps. This fact may give the Government an extra excuse for counselling patience until the next report from the Molony committee. 

In the first place our business is foreign policy, and it is the business of the Presidential leadership and his appointees in the Department to consider the domestic political aspects of a problem. Mr Truman emphasised this point by saying, ‘You fellows in the Department of State don’t know much about domestic politics’.

Here a passage of text is picked up by anaphoric reference, as in the case of text reference by means of this and that on their own; but fact nouns add a classification and often an assessment (which may be supported by post-Deictics or Epithets) of the discursive antecedent:

There is a subdued aspect of the current political voices: with all the tension generated by the electoral process, it is only a means to an end. The end actually is the transformation in the quality of lives of the people. We must never lose focus of this as an issue. This obvious point can certainly not be over-emphasised.

Blogger Comments:

Here again Matthiessen misrepresents Halliday's model of cohesive reference.  Like nouns of projection, fact nouns do not refer anaphorically or cataphorically.  It is the reference item that precedes the noun (the, this etc.) that refers.  The absence of a reference function of a noun can be made obvious by simply removing the reference item that precedes it, and trying to identify the referent of the noun.

To be clear, any classification or assessment of a fact noun is made through nominal group structure, and is distinct from the system of cohesive reference.

Facts: Assessment

Halliday & Matthiessen (2014: 537-8):

The ‘fact’ noun serving as Head/Thing of the nominal group may thus embody a form of assessment of the projected clauses serving as Postmodifier/Qualifier – assessment involving modalisation. Premodifiers may provide further assessments of the projected clause, either as attitudinal Epithet (e.g. painful, interesting, obvious; cf. good in good chance above) or as post-Deictic (e.g. alleged), e.g.

‘The U.S. government has to come to terms with the painful fact [[that the good old days [[when it could just borrow its way out of messes of its own making]] are finally gone]],’ Xinhua wrote. 

It’s an interesting fact [[[that, <<compared with other countries, >> Australians are not very heavy drinkers]]]. 

The alledged (sic) fact [[that a motor breakes (sic) more easily under moderated modifications]] isn’t a symptom of anything going wrong with it –# just that the stock setup/tune is closer to its limits than it used to be. 

No one would like to contend the blatantly obvious fact [[that thought and consciounsness (sic) do not fall into the category of material objects according to the current definitions of matter]].

These often correspond to comment Adjuncts in ‘declarative’ clauses (e.g. interestingly, Australians are not ...); but unlike statements realised by ‘declarative’ clauses, facts are not open to direct challenge in dialogic interaction.

Facts: Cases vs Chances vs Proofs

Halliday & Matthiessen (2014: 537):

The first three differ in terms of modality of the subtype modalisation:

(1) cases (nouns of simple fact) relate to ordinary non-modalised propositions ‘it is (the case) that ... ’ 

(2) chances (nouns of modality) relate to modalised propositions ‘it may be (the case) that ... ’ 

(3) proofs (nouns of indication) relate to propositions with indications, which are equivalent to caused modalities, ‘this proves/implies (i.e. makes it certain/probable) that ... ’

The Four Sub-Classes Of Fact Noun

Halliday & Matthiessen (2014: 537, 536):

There are four sub-classes of fact noun: (1) cases, (2) chances, (3) proofs and (4) needs. The first three go with embedded projections [typo: propositions] whereas the last goes with embedded proposals; see Table 7-26 for examples of nouns.

Facts: Prepackaged Projections

Halliday & Matthiessen (2014: 537):

There is one other type of projection, where the projected clause is not being projected by a verbal or mental process with Sayer or Senser, or by a verbal or mental process noun in a metaphorical nominal group, but comes, as it were, ready packaged in projected form. We refer to this type as a fact. For example:

The irony is even further compounded by the fact [[that while every people must have those [[[who say, || ‘Here I stand,’]]] the fact is [[[Okonkwo loses a child to the forces of Christianity, || and Ezeulu loses his community to Christianity]]] ]].

The fact [[that fourteen of the original eighteen Julian Ashton nudes still decorate the Marble Bar’s walls]], perchance contributed to the bar winning Australian Playboy’s survey for Best Bar in Australia in 1986.

The fact [[that Lear never even alluded to that at the end]] is a sign [[that he didn’t learn very much through the course of the play]].

He’s trapped by the fact [[that the river flows south]].

Here a ‘fact’ noun serves as the Head/Thing of a nominal group with a projected clause as Postmodifier/Qualifier. Consider That Caesar was dead was obvious to all. Here that Caesar was dead is certainly a projection; but there is no process of saying or thinking which projects it. Its status is simply that of a fact; and it can indeed function as Qualifier to the noun fact, e.g. the fact that Caesar was dead was obvious to all.

The Projecting Environments Of Projected Propositions And Proposals

Halliday & Matthiessen (2014: 536):

Thus verbal processes, and mental: cognitive processes, project in the indicative mode (propositions), while verbal processes, and mental: desiderative processes, project in the imperative mode (proposals). The projecting environment may be a verbal or mental process clause, or a (metaphorical) nominal group with a verbal or mental process noun (locution or idea) as its Head.

The Subject Of Non-Finite Embedded Projections

Halliday & Matthiessen (2014: 536):

Where the projected clause is non-finite the Subject can be presupposed from the primary clause provided it is the participant that is actually doing the projecting – Senser or (more rarely) Sayer. So the thought of being a queen (encouraged her), her desire to be a queen ... , her assertion of being a queen ... , where ‘she’ is doing the thinking, etc.; but the news of her being a queen (proclaimed by someone else), the thought of her being a queen (in someone else’s mind), and so on. These correspond to the non-finite forms with hypotaxis: she wanted to be a queen, they wanted her to be a queen. In the finite forms, of course, the Subject is always made explicit.

An Embedded Projection Defines Its Noun Of Projection

Halliday & Matthiessen (2014: 535):

… the noun is the name of a locution or an idea, and the clause that it projects serves to define it in exactly the same way that a ‘restrictive’ relative clause defines the noun that is expanded by it. Hence any noun that belongs to a projecting class may be defined (restricted) in either of these two ways, either by projection (e.g. the thought [[that she might one day be a queen]]) or by expansion (e.g. the thought [[that came into her mind]]). This leads to ambiguities such as the report [[that he was submitting]].

Nouns Of Projection

Halliday & Matthiessen (2014: 535, 536):

Nouns that project belong to clearly defined classes, verbal process nouns (locutions) and mental process nouns (ideas); they correspond rather closely to, and in many instances are derived from, the verbs used in the projecting clause, especially the reporting ones. Some of the principal nouns of projection are set out in Table 7-26.

The nature of the realisation of the embedded clause depends on the speech functional subcategory:

(I) Propositions

(a) stating: projected clause either (i) finite, that + indirect indicative, or (ii) non-finite, of + imperfective 

(b) questioning: projected clause either (i) finite, if/whether or WH- + indirect indicative, or (ii) nonfinite, whether or WH- + to + perfective

(II) Proposals

(a) offering (incl. suggesting): projected clause either (1) non-finite, to + perfective or of + imperfective, or (ii) finite, future indirect indicative 

(b) commanding: projected clause either (i) non-finite, to + perfective, or (ii) finite, modulated or future indirect indicative

Embedded Projections In The Creation Of Discourse

Halliday & Matthiessen (2014: 534-5):

The fact that the projected clause is embedded as the Qualifier in a nominal group means that it can occur in a range of grammatical environments not open to non-embedded, tactically related projected clauses. This is important in the creation of discourse; one of the central uses of nominal groups with embedded projections is in the representation of arguments, as in newspaper reports and scientific discourse:

There is bitter opposition to his proposal [[that Palestinians renounce their demand [[for more than three million refugees to return to areas inside Israel that were abandoned in the 1948 war]] ]]. 

Israelis have rejected Mr Clinton’s proposal [[that they give up control of the Temple Mount in Jerusalem’s walled Old City, the holiest place in Judaism and the third most sacred in Islam]].

Boyle’s tentative suggestion [[that heat was simply motion]] was apparently not accepted by Stahl, || or perhaps it was unknown to him.

Here proposals and demands are opposed, renounced and rejected. The contribution to the creation of discourse is further enhanced by the fact that such nouns of projection can be used anaphorically to refer back to propositions and proposals already established in the discourse:

The Labour Party opposed Thor missiles, because, he said, they were out of date and vulnerable and would attract enemy action. That argument did not apply to the Polaris submarine.

The cohesive effect is similar to that created by text references achieved by means of this, that, it:

The talks lasted for three hours. This was a surprise, for they had only been scheduled to last two hours.

(cf. it was a surprise that the talks lasted for three hours with a fact clause; but nouns of projection make it possible to construe the class of projection explicitly.

Blogger Comments:

To be clear, here Matthiessen seriously misrepresents Halliday's model of cohesive reference.  It is not the noun of projection (argument) that makes anaphoric reference, but the reference item that precedes it (that). It would appear that Matthiessen has been fooled by the misunderstanding of Halliday in Martin (1992).  For evidence that Martin misunderstands the principles on which reference is theorised, see any of the 153 posts here.

All Embedded Projections Are Metaphorical

Halliday & Matthiessen (2014: 534):

The structure of a nominal group with an embedded projection is shown in Figure 7-22, part [i].

Such instances are still projections; but here the projecting element is the noun that is functioning as Thing, in this case assertion. … such instances of projection are all metaphorical: a projecting sequence is realised congruently as a clause nexus of projection – part [ii] of Figure 7-22, but it may alternatively be realised metaphorically as a nominal group – part [i] of Figure 7-22. When we align them as in Figure 7-22, we see how the nominal group construction with an embedded projection clause is agnate with a clause nexus of projection: the nominal group is a metaphorical, nominalised version of the clause nexus; and the noun assertion serving as Head/Thing is in fact a nominalised variant of the verb assert serving as Process in the agnate clause. The congruent Sayer may be left out in the nominal group; or it may be represented either as the Deictic (their assertion that ...) or as a Qualifier (the assertion by the government that ...). Part of the rhetorical power of the metaphorical group is the potential for leaving the Sayer unspecified.

Embedded Locutions And Ideas

Halliday & Matthiessen (2014: 533-4):

Like the three types of expansion, both locutions and ideas can be embedded. Besides entering into paratactic and hypotactic clause nexuses, they can be ‘rankshifted’ to function as Qualifiers within a nominal group, as in:

Leaders of both a publicly-funded project and a competing private company issued statements Friday [[that they jointly would announce the status of their work on Monday]].

I was very intrigued by your take on Huck Finn in that piece, and your argument [[that the great American novel of that century was Uncle Tom’s Cabin]].

To what extent do you buy into the belief [[[that if the individual becomes enlightened, || that adds to the betterment of the universe in and of itself]]]?

AT&T’s stock slid 14 percent Tuesday as the company issued its first profit warning under chief executive C. Michael Armstrong, fuelling worries about [[whether his radical remake of the nation’s largest long-distance company will succeed]]. 

The man was impressive in some ways, Oxford educated, very twenties British bohemian, a great dancer and seducer of women, who suppressed his wife’s desire [[to be a ‘real’ archaeologist]] and whose own career really was a joke up until his early death from a sudden illness.

The Intonation Pattern Of Free Indirect Speech

Halliday & Matthiessen (2014: 532):

The intonation pattern of free indirect speech is still further anomalous, since it follows that of quoting and not that of reporting: the projected clause takes the intonation that it would have had if quoted (that is, identical with its straight, unprojected form), and the projecting clause follows it as a ‘tail’. This is because the projected clause still has the status of an independent speech act.

Direct, Free Indirect And Indirect Speech

Halliday & Matthiessen (2014: 532, 533):

To accommodate free indirect speech in our account, we thus need to expand Table 7-18 by dissociating the quote vs. report variable from the parataxis vs. hypotaxis one: see Table 7-25. As the table shows, free indirect speech can be projected both verbally and mentally, and includes both propositions and proposals – everything, in fact, that can be both quoted and reported (thus excluding minor speech functions since they can only be quoted).

Free Indirect Speech

Halliday & Matthiessen (2014: 532, 532n):

… there is another mode of projection which is sometimes described as ‘intermediate between direct and indirect speech’, namely free indirect speech:³²

Quoted (‘direct’)           ‘Am I dreaming?’ Jill wondered
‘Free indirect’                Was she dreaming, Jill wondered
Reported (‘indirect’)      Jill wondered if she was dreaming

Strictly speaking it is not so much intermediate as a blend: it has some of the features of each of the other two types. The structure is paratactic, so the projected clause has the form of an independent clause retaining the mood of the quoted form; but it is a report and not a quote, so time and person reference are shifted – was she not am I. This is another example of the semogenic principle whereby the system fills up a slot it has created for itself.

∞ 

³² ‘Free indirect speech’ encompasses a range of different feature combinations; it is a projection ‘space’ rather than a single invariant pattern. The account given here represents it in its prototypical form.

Referring To Quoted Ideas vs Substituting For Reported Ideas

Halliday & Matthiessen (2014: 531):

With ‘mental’ process clauses the picture is more complex, since the reference form that tends to be associated with certainty and the substitute so with uncertainty; the principle is actually the same, but it is operating in a different environment (cf. the different senses of thought in quoting and reporting, referred to above). The principle is that a substitute does not refer; it simply harks back. It thus has the general semantic property of implying, and so excluding, possible alternatives; cf. the nominal substitute one as in a big one, meaning ‘there are also small ones, and I don’t mean those’. This is why so, which is a clause substitute, has the general sense of ‘non-real’, by contrast with what is ‘real’; besides (i) projection, where it signifies what is asserted or postulated, it is used in two other contexts: (ii) hypothetical, as opposed to actual, and (iii) possible, as opposed to certain. Hence:

(i) I think so          but      I know [that]         not      I know so
(ii) if so                 but      because of that      not      because so
(iii) perhaps so      but      certainly                not      certainly so

Referring To Quoted Locutions vs Substituting For Reported Locutions

Halliday & Matthiessen (2014: 530-1):

Compare:

The sky is about to fall. (i) – Who said that? (ii) – Who said so?

 It is clear that both that and so stand for something that is projected, as shown by the verb said. In (i)  this projected element is being treated as a quote: ‘who produced that verbal act?’ – hence we can ask who said that? if we want to identify a speaker from among a crowd, like a teacher finding out who was talking in class. In (ii), on the other hand, the expression the sky is about to fall is being treated not as anybody’s verbal act but as a text; the meaning is ‘who affirmed that that was the case?’, with the implication that the contrary is conceivable.

In ‘verbal’ process clauses, therefore, he said that simply attests his production of the wording, whereas he said so raises the issue of whether what he said is in fact the case.

Referring To Quotes vs Substituting For Reports

Halliday & Matthiessen (2014: 530):

There are different ways of referring back to what is quoted and what is reported. Typically a reference item, usually that, is used to pick up a quoted passage, while a substitute, so/not, is used with a report. For example,

She said, ‘I can’t do it.’– Did she really say that?
She said she couldn’t do it.– Did she really say so?

This is because the act of quoting implies a prior referent, some actual occasion that can then be referred back to, whereas in reporting there is nothing but the reported text. This explains the difference in meaning between I don’t believe that ‘I do not accept that assertion as valid’ and I don’t believe so ‘in my opinion such is not the case’.

The Semantic Inequivalence Of Direct And Indirect Speech

Halliday & Matthiessen (2014: 530):

Traditional school exercises of the kind ‘turn into direct/indirect speech’ suggest that the two always fully match. This is true lexicogrammatically, in that it is always possible to find an equivalent – although not always a unique one: given Mary said she had seen it, the quoted equivalent might be I have seen it, I had seen it or I saw it, or she (someone else) has seen it, etc. But it is not true as a general statement about usage. Semantically the two do not exactly match, and there are many instances where it does not make sense to replace one by the other. Note, for example, Alice thought that that was the jury-box, where we should have to change Alice thought to something like Alice said to herself in order to avoid the sense of ‘held the opinion’, which is the natural interpretation of a verb of thinking when it is projecting by hypotaxis.

Reported Imperative (‘Indirect Command’)

Halliday & Matthiessen (2014: 530):

With the imperative the relationship is less clear. We noted in Chapter 4 that the imperative is a somewhat indeterminate category, having some features of a finite and some features of a non-finite clause. Similarly the category of reported imperative (‘indirect command’) is not very clearly defined. But non-finite clauses with to, following a verb such as tell or order, can be interpreted as reported proposals. They likewise display the properties of ‘indirect speech’, although without sequence of tenses, since the verb does not select for tense. For example,

‘I know this trick of yours.’ She said || she knew that trick of his.
‘Can you come tomorrow?’ He asked || if she could come the next day.
‘Why isn’t John here?’ She wondered || why John wasn’t there.
‘Help yourselves.’ He told them || to help themselves.
‘We must leave to-night.’ She said || they had to leave that night.

The Shift Of Mood In Reported Interrogatives

Halliday & Matthiessen (2014: 530):

If the reported clause is interrogative it typically shifts into the declarative; the declarative is the unmarked mood, and is used in all clauses that do not select for mood independently, including all dependent clauses. A yes/no interrogative becomes declarative, introduced by if or whether (he asked ‘is she coming at noon?’ : he asked whether she was coming at noon); a WH- interrogative becomes declarative with the WH- element remaining at the front (he asked ‘when is she coming?’ : he asked when she was coming).

The Shift Of Deixis And Tense In Reported Propositions

Halliday & Matthiessen (2014: 529):

As we have seen, a reported proposition typically takes on a set of related features collectively known as ‘indirect speech’. What happens is that all deictic elements are shifted away from reference to the speech situation: personals away from first and second person (speaker and listener) to third, and demonstratives away from near (here-&-now) to remote. A part of this effect is the ‘sequence of tenses’: if the verb in the reporting clause has ‘past’ as its primary tense, then typically each verb in the reported clause has its finite element in the corresponding System II (‘sequent’) form: see Table 7-24.

In other words, an additional ‘past’ feature is introduced at the Finite element in the mood structure of the projected clause. The use of the sequent form is not obligatory; it is less likely in a clause stating a general proposition, for example they said they close at weekends. But overall it is the unmarked choice in the environment in question.

Quoting And Reporting As Modes Of Projection

Halliday & Matthiessen (2014: 528):

Quoting and reporting are not simply formal variants; they differ in meaning. The difference between them derives from the general semantic distinction between parataxis and hypotaxis, as it applies in the particular context of projecting. In quoting, the projected element has independent status; it is thus more immediate and lifelike, and this effect is enhanced by the orientation of the deixis, which is that of drama not that of narrative. Quoting is particularly associated with certain narrative registers, fictional and personal; it is used not only for sayings but also for thoughts, including not only first-person thoughts, as in

... and watching that trial wondering whether in fact he was innocent or not and I couldn’t make up my mind, after a while I thought ‘No, I’m sure he’s guilty’.

but also third-person thoughts projected by an omniscient narrator, as in

‘And that’s the jury-box,’ thought Alice.

So after about two hours he thought ‘Well they’re not coming back’ and he started hitchhiking.

Reporting, on the other hand, presents the projected element as dependent. It still gives some indication of mood, but in a form which precludes it from functioning as a move in an exchange; the mood is projected, not straight. And the speaker makes no claim to be abiding by the wording.