The word "grammatical" tends to be used in a very loose sense in this context. It is usual to illustrate the grammatical function by inventing sentences which when written are ambiguous, and whose ambiguity can only be removed by using differences of intonation. A typical example is the sentence 'Those who sold quickly made a profit'. This can be said in at least two different ways:

i) 'Those who 'sold vquickly I ,made a profit

ii) 'Those who vsold I ,quickly .made a profit

The difference caused by the placement of the tone-unit boundary is seen to be equivalent to giving two different paraphrases of the sentences, as in:

i) A profit was made by those who sold quickly.

ii) A profit was quickly made by those who sold.

Let us look further at the role of tone-unit boundaries, and the link between the tone- unit and units of grammar. There is a strong tendency for tone-unit boundaries to occur at boundaries between grammatical units of higher order than words; it is extremely common to find a tone-unit boundary at a sentence boundary, as in:

I 'wont have any /tea. I 'dont like it

In sentences with a more complex structure, tone-unit boundaries are often found at phrase and clause boundaries as well, as in:

In vFrance | where .farms .tend to be vsmaller | the 'subsidies are 'more important

It is very unusual to find a tone-unit boundary at a place where the only grammatical boundary is a boundary between words. It would, for example, sound distinctly odd to have a tone-unit boundary between an article and a following noun, or between auxiliary and main verbs if they are adjacent (although we may, on occasions, hesitate or pause in such places within a tone-unit; it is interesting to note that some people who do a lot of arguing and debating, notably politicians and philosophers, develop the skill of pausing for breath in such internationally unlikely places because they are less likely to be interrupted than if they pause at the end of a sentence). Tone-unit boundary placement can, then, indicate grammatical structure to the listener and we can find minimal pairs such as the following:

i) The Con'servatives who vlike the pro-posal I are pleased

ii) The Convservatives I who vlike the pro-posal I are pleased

The intonation makes clear the difference between (i) "restrictive" and (ii) "non-restrictive" relative clauses: (i) implies that only some Conservatives like the proposal, while (ii) implies that all the Conservatives like it.

Another component of intonation that can be said to have grammatical significance is the choice of tone on the tonic syllable. One example that is very familiar is the use of a rising tone with questions. Many languages have the possibility of changing a statement into a question simply by changing the tone from falling to rising. This is, in fact, not used very much by itself in the variety of English being described here, where questions are usually grammatically marked. The sentence 'The price is going up' can be said as a statement like this:

The price is going -up

(the tonic stress could equally well be on 'up'). It would be quite acceptable in some dialects of English (e.g. many varieties of American English) to ask a question like this:

(Why do you want to buy it now?) The /price is going -up

But speakers in Britain would be more likely to ask the question like this:

(Why do you want to buy it now?) 'Is the /price going -up

It is by no means true that a rising tone is always used for questions in English; it is quite usual, for example, to use a falling tone with questions beginning with one of the "wh-question-words" like 'what', 'which', 'when', etc. Here are two examples with typical intonations, where (i) does not start with a "wh word" and has a rising tone and (ii) begins with 'where' and has a falling tone.

i) 'Did you 'park the /car

ii) 'Where did you 'park the car

However, the fall in (ii) is certainly not obligatory, and a rise is quite often heard in such a question. A fall is also possible in (i).

The intonation of question-tags (e.g. 'isn't it', 'can't he', 'should she', 'won't they', etc.) is often quoted as a case of a difference in meaning being due to the difference between falling and rising tone. In the following example, the question-tag is 'aren't they'; when it has a falling tone, as in (i), the implication is said to be that the speaker is comparatively certain that the information is correct, and simply expects the listener to provide confirmation, while the rising tone in (ii) is said to indicate a lesser degree of certainty, so that the question-tag functions more like a request for information.

i) They 'are 'coming on Tuesday I arent they

ii) They 'are 'coming on Tuesday I /arent they

The difference illustrated here could reasonably be said to be as much attitudinal as grammatical. Certainly there is overlap between these two functions.