Questions in spoken language

Spoken language is usually seen as being more interactive than written language. As speakers, we address each other directly (Hey guys), indicate our attention to each other (Mmm), and respond to each others’ comments (Really?, You didn’t!). These are all examples of interactive features.

Another interactive feature associated with spoken language is question–answer sequences. In this investigation we will explore this feature, using data from ICE-GB (our corpus, or database of real language).

In ICE-GB we can do an automatic search for interrogative clauses, a type of clause typically used to ask questions. Interrogative clauses have particular grammatical features involving a special word order (Subject-verb inversion) and/or the presence of a question word such as who, what, where. Here are a couple of examples from the corpus:

• Can you see that _{[S1A-087 #183]}
• When did you get married _{[S1A-056 #230]}

Note that we don’t see the expected question marks in these examples, as final punctuation is not used for spoken language units in the corpus.

Spoken vs. written English

The points above give us a starting point for our investigations. We know that interrogative clauses are typically used to ask questions, and that interactive features like questions are typically associated with spoken language. This might lead us to ask the following question:

• Are interrogative clauses more frequent in spoken than in written English?

Or, instead of wording this as a question, we can word it as a hypothesis, an idea which is to be tested against the data:

• Interrogative clauses are more frequent in spoken than in written English.

Spoken vs. written: Step 1

Step 1. To test our hypothesis, we start by looking for main clauses which are interrogative in ICE-GB. Our search finds the following:

This result seems to show a very striking difference. It looks as though there are around seven times as many interrogative clauses in the spoken material as in the written. But is it correct to compare the numbers directly in this way?

There are a couple of reasons why this is not a fair comparison:  

• The amounts of spoken and written material in the corpus are not the same (about 60% is spoken and 40% written).
• Main clauses in speech tend to be shorter than those in writing. This means that there are likely to be more main clauses per thousand words in spoken transcripts than in written text – so there are more opportunities to use an interrogative main clause in speech.

To do a proper comparison, we need to:

• find out the total number of main clauses in the spoken and written material, and then
• work out what proportion of these are interrogative.

Spoken vs. written: Step 2

Step 2. We search for all main clauses in ICE-GB, and add these figures to our table:

Spoken vs. written: Step 3

Step 3. We divide the number of interrogative main clauses by the total number of main clauses to find out what percentage of main clauses are interrogative.

Spoken vs. written: discussion

• Our results show that 11% of clauses are interrogative in the spoken material compared to 3% in the written material. This is nearly four times as many – instead of seven times as many, as it appeared at Step 1.
• This is still a striking difference. Now that we have made a proper comparison, we can confidently state that the ICE-GB data supports our original hypothesis (that interrogative clauses are more frequent in spoken than in written English).
• If we wanted to be even more scientific, we might perform a statistical test, but we have a lot of data so we can be quite confident in its reliability.

Dialogue vs. monologue

We can take our investigation further by comparing different types of spoken data. There are two main groupings of spoken genres in ICE-GB: dialogue and monologue. (We leave aside a smaller category of ‘mixed’ type.)

• The dialogue material includes genres such as private conversations, business transactions and parliamentary debates.
• The monologues include genres such as speeches (scripted and unscripted), sports commentaries and teaching demonstrations.

Which of the two groupings, dialogue or monologue, do you think will have the highest frequency of interrogatives? Write down your idea now in the form of a hypothesis.

If we follow the three steps we used for written vs. spoken data, we obtain the following table (with the smaller ‘mixed’ spoken category removed from the total).

• The percentage of interrogatives is very much higher in the dialogue material than in the monologues. Does this support your hypothesis?
• The figure for monologues (2%) is even lower than that for the written material we looked at earlier (3%). Do you find this surprising? Do we need to reconsider claims that spoken language is more interactive than written language?

Comparing dialogue genres

Let’s look more closely at the different genres of dialogue in ICE-GB. There are eight genres:

• private conversations
• telephone calls
• broadcast discussions (on TV or radio)
• broadcast interviews (on TV or radio)
• business transactions
• classroom lessons (mainly university seminars)
• legal cross-examinations
• parliamentary debates

Which genre do you think will have the highest proportion of interrogative clauses? And which the lowest? Write down your ideas as hypotheses now.

The table below shows what we find in ICE-GB (with the category names abbreviated). You may need to scroll down to see the full results for the eight categories and the total.

Results for dialogue genres

Review of findings

The percentages of interrogatives vary considerably among these genres.

• Which have the highest and lowest percentages of interrogatives? Do the findings agree with your predictions?
• What possible explanations might there be for these patterns? How might you go about testing them?

Looking at a spoken extract

One way to investigate further is to look in detail at particular extracts. Individual extracts themselves vary in the frequency of interrogatives, so we could choose some with especially high frequencies.

In the extract below, 46% of the main clauses are interrogatives. It is taken from a classroom lesson at a university. Here are some points to help you understand the layout of the extract:

• There are three speakers: A  is the lecturer, and B and C are students. Their contributions are highlighted in different colours (B in blue and C in green).
• Smaller sections of highlighting inside square brackets indicate overlapping (i.e. stretches where speakers are talking at the same time).
• The interrogative main clauses are in a pink font, so you can find them easily.
• The markers (.) and (1) indicate shorter and longer pauses, and question marks indicate unclear speech.

Classroom lesson extract

Questions for discussion

• What do you notice about who is using the interrogative clauses in this extract?
• Why do you think there are so many interrogatives in the extract? What is the purpose behind the questions being asked? Does the questioner want to find out information that he or she doesn’t know?
• Can you identify some tag questions among the interrogative clauses? How are these being used in the interaction?
• The first contribution from B is