The texts studied were transcripts of budget speeches given to Parliament. They comprise two given by the then Chancellor of the Exchequer, Kenneth Clarke, to the British House of Commons in 1994 and 1995, and two given by the current Chancellor, Gordon Brown, in 2001 and 2002. I also looked at the budget speech given by the Indian Finance Minister to that country’s Parliament in 2002.

Each budget had its own significant phrases that were revealed by small word cluster searches. These provide an indication of the preoccupations and overall tone of each speech. The most frequent are function word clusters such as ‘of the’, so the table shows a selection of lexically important clusters.

This overview reveals differences in presentation: Clarke uses the traditional deferential form ‘My Right Honourable Friend’ to refer to colleagues in 1994 but the next year he calls them by name. There is a shift from first person plural to singular across time: Clarke uses ‘we’ and ‘the government’ to portray himself as part of a collective decision, but Brown tends to use ‘I’. This may be symptomatic of a recent trend away from Cabinet government towards a scenario in which a handful of important ministers vie for credit and power. Brown’s use of modal verbs also portrays a more proactive and forward-thinking approach. He uses ‘I have’ and ‘we have’ to review the previous year’s achievements much less frequently than Clarke, and there is a greater incidence of verbs indicating future action such as ‘can’, ‘will’ and ‘propose.’ It is a combination of the move towards personal identification with the policies and the need to be seen to act that leads Brown to use the phrase ‘I have decided’ nine times during the speech. The Indian Finance Minister also speaks in the first person and makes a total of 104 proposals in this form. This may be a matter of oratory style that differs between the two countries and between native and non-native speakers. However, it may be that he will find it much more difficult to get his budget ratified than Brown will his, so anything put forward really is of the tentative nature of ‘I propose to...’ rather than ‘I will...’

The frequent lexical clusters vary from budget to budget and highlight the main economic concerns at that time. A concordance of the phrase ‘public spending’ in the 1994 budget shows that it collocates nine times (out of twenty) with ‘control’ or similar words and five times with ‘reduce’ or ‘cut’. In 2001 the phrase does not appear at all, and talk is of increasing ‘spending’. This reflects the differing priorities of the two administrations: the Conservative government trying to lift the country out of recession, and the Labour government trying to fulfil election pledges to improve public services. Hence also the frequent mentions of ‘the NHS’ in 2002’s budget against just seven in the two Clarke budgets.

A more curious example is the use of ‘long term’ and ‘long-term’. Three of the four times it comes up in 1994 it refers to long-term care of the elderly or disabled, and in 1995 to long-term unemployment. Only twice does Clarke use it in a more abstract sense of looking to the future. Brown, on the other hand, uses it constantly in this sense, talking of ‘long-term objectives’ and ‘a secure long term financial foundation’. There may be an element of spin in this: it is highly unlikely that Clarke thought only in the short term as Chancellor, but Brown and his speech-writers want to make sure the electorate trust them to handle the economy across a period of time.

When comparing wordlists drawn from the British budget speeches, a major difference soon becomes apparent in their reference to families and children. When all the lemmas of ‘family’ and ‘child’ had been identified, the effect was even stronger.

Clearly, these words are central to the policies of Brown’s budgets and this can be confirmed by comparing concordances from the two decades of British economic strategy represented here. The 1994 budget mentions ‘couples with children’ and ‘childless couples’ a small number of times. The 1995 budget is different: it talks of ‘their [i.e. voter’s] children’ when referring to the impact of actual policies and ‘our children’ in more aspirational passages that look to the future. However, there are still only eight instances overall. By contrast ‘child’ or related words occur 53 times in the 2001 speech and 50 times in the 2002 speech, with 0.62% and 0.54% frequency respectively. This commonness is explained when they are concordanced: Brown regularly uses the plural as a noun and the singular even more frequently as a modifier, especially in ‘child tax credit’, ‘child poverty’ and ‘child benefit’.

‘Family’ appears eight times in the 1994 speech, each time in the phrase ‘family credit’. In 1995 there is more general consideration of families and the word often appears in more ideologically tinted propositions such as ‘These are the people of Britain who are hard-working and take care of themselves and their families’. In Brown’s 2001 budget ‘family’ and its lemmas collocate 26 times (out of 42) with ‘tax’, four times in the phrase ‘working families tax credit’, but also frequently in terms of revision of the tax system for the benefit of families. The 2002 budget is even more centred on ‘family’, with 50 instances making up 0.54% of the text. This is very high for a content (rather than function) word, meaning that family must be an important topic in the speech. The key paragraph would appear to be this:

Later there are again mentions of ‘working families tax credit’ which becomes incorporated into this wider ‘family friendly tax system’. There is renewed concern for ‘the poorest families’ (two instances) and new terminology is introduced in the form of ‘single-’ or ‘two-earner families.’ There is also a tendency to use ‘family’ as a modifier in phrases that imply a more domestic kind of budget: ‘family life’ and ‘family income’.

The Indian budget does not deal much with families or children. In most cases the families mentioned belong to specific groups such as ‘defence forces and their families’. The only occasion when it shows some similarities to the British budgets is a concern for nursing mothers in below poverty line families. Despite there being a department for Women and Child Development, it receives only a very short paragraph in the budget which notes problems with malnutrition. The much lower frequency of ‘family’ and ‘child’ in the budget reflects India’s different stage of economic development and the fact that its government is more worried about other things.

Overall, this small exploration of the use of ‘family’, ‘child’ and related words has illuminated the shift in political attitudes in this area over the last decade. Clarke’s 1994 budget is perhaps the last one in which the family unit and people’s personal financial situations come second to macro-economic considerations. In 1995 these words were used in an explicitly aspirational and ideological way; by the time Gordon Brown gave his budget speeches the ideology is more subtly delivered, but the emphasis on family and children even greater.

I think Wordsmith did enable me to study the texts in ways that would not otherwise have been possible. It was the Keyword program that flagged up the big difference between the texts in their use of ‘child’ and ‘family’ and from that I could go on to use Concord to examine the instances of these words in more detail. Without Wordsmith I may have noticed the focus on ‘family’ and ‘child’ in Brown’s budgets, but it would have been a long and tedious task to systematically compare them to the other budgets and to find patterns in their use. Similarly, I found making a cluster Wordlist to be a quick way of picking out the major considerations in each budget.

The collocation tool in Concord was quite frustrating because it has the potential to be very useful but, because it cannot recognise lemmas, it often fails; for example, in the 2002 budget I noticed that ‘children’ collocates once with ‘responsibility’ and once with ‘responsibilities’, but the program could not pick up that they were essentially the same word. Keyword was also sometimes misleading. It told me that Brown uses ‘pounds’, ‘pence’, ‘million’, ‘billion’ and ‘per cent’ much more than Clarke and the Indian budget. This is simply because of differences in notation (‘£6m’ versus ‘6 million pounds’ and ‘25%’ versus ’25 per cent’) and in currency (the Indian budget is naturally in rupees). To overcome this problem I would have to go through all the texts and standardise them – a time-consuming job. Although the Indian budget was essentially similar to the British ones, there were enough small differences in style and content to render Wordsmith’s comparisons virtually useless; this is why I have not mentioned it much in my analyses above. This is a definite limitation of the program: it is effective in comparing like with like or two completely different texts but there is a range within which no meaningful conclusions can be drawn.

There were other investigations I was able to do in the time which I have not included here, such as looking at the use of the politically charged phrases ‘single mother’ or ‘lone parent’. Given more time, I would probably have looked at words with a left-right bias such as ‘public’, ‘private’ and ‘capital’ to see how the Labour budgets differ from the Conservative ones. A large-scale analysis of this kind might shed light on accusations that ‘New’ Labour simply stole Tory ideas and language to get elected. I might also have found the texts of budgets from other decades and countries and compared them.

Using Wordsmith has been a unusual introduction to the world of corpus linguistics because I did not use any corpora. I expect I would have looked at completely different aspects of the texts if I had. It was useful to have empirical evidence to back up my arguments, and I am pleased that the topics I investigated were guided by the program, rather than being my own idea. However, it is clear from my analyses above that I have still imposed my own interpretation on the statistics. In one way this is necessary as Wordsmith has no knowledge of political history and cannot do it for me, but it still undermines the empirical aims of computer-aided stylistics. Perhaps one day a successor to Wordsmith will incorporate encyclopaedic knowledge and be able to come up with its own ideas as to what its statistical findings mean; until then, Wordsmith and computer analysis in general are just another set of tools for the stylistician to use – with varying degrees of success.

Essay index! -or- Take me home!