Help!

(1) He helped his brother build a wall

    (2) He helped his brother to build a wall

Which of these two sentences would you say: (1) or (2)? Most English speakers use both, but when do they choose one rather than the other, and why?

Previous researchers have found that people use the “bare” infinitive (as in help build) more often than the full infinitive (help to build) when they are speaking than when they are writing, and more often in informal situations.  However, Arne Lohmann discovered that things are a bit more complicated than that. He analysed 1718 examples of help with a following infinitive – both bare and full – in the British National Corpus.

There was a complex set of findings for spoken and written English, though overall different genres of spoken English did seem to favour the bare infinitive more than non-fiction written English. However, what influenced the use of the bare infinitive more than anything else was when there was a second full infinitive after HELP: so, with a sentence like (3), people were more likely to say help the candidate relax than help the candidate to relax.

(3) your job as an interviewer is to help the candidate (to) relax

Also relevant, though less so, was the number of words between HELP and the infinitive. The bare infinitive was more likely in a sentence like (4), where there are three words (the home decorator) between HELP and the infinitive.

(4) the automatic wallpaper pasting machine helps the home decorator to achieve really professional results

It was less likely in a sentence like (5), where there is only one word (me).

(5) I don’t mind talking about it because it helps me get over what happened.

The more words between HELP and the infinitive, the more likely was the full infinitive, probably because as sentences get more complex, including to makes the grammatical structure more explicit.

A further relevant factor was when the subject of HELP was human: so, the fact that the subject of (3) is an automatic wallpaper pasting machine may account for the use of the full infinitive in that sentence. Perhaps we see a human subject as more involved in the helping event, and this small conceptual distance between the subject and the verb is then matched by a small linguistic distance between HELP and the meaning of the infinitive.

Why should a second infinitive have such a strong influence on saying help build rather than help to build? Lohmann suggests that this reflects our system for speech production. Current theories assume that words in our mind are organized in the same way as neural networks, with words roughly corresponding to neural nodes, with links between them. Like neurons, when a node has been activated it needs a recovery phase. Speakers therefore avoid repeating a similar or identical form, as this would use the same node again in the network.

Like all research, finding a plausible answer to this small question raises many other questions that now need to be answered. What caused English speakers to drop to after HELP in the first place? Do they drop it only with HELP, or are there other verbs that can have both a bare infinitive and a full infinitive after them? Is this more common in American English than British English? And what about other types of grammatical construction, as in all we have to do is (to) wait?

_________________________________________________________

Lohmann, Arne (2011). Help vs help to: a multifactorial, mixed-effects account of infinitive marker omission. English Language and Linguistics 15 (3): 499-521.

doi: 10.1017/S1360674311000141

This summary was written by Jenny Cheshire

Who's the expert?

The linguist?

Who decides where one regional dialect stops and another begins? And who decides what are the distinctive features of a dialect? Barbara Johnstone’s research in Pittsburgh, USA, finds that linguists and local people tend to give different answers to these questions.

Local people talk about “Pittsburghese” and are interested especially in words that they think are specific to their city. They assume that the features of Pittsburghese are used only in the Pittsburgh area, and that everyone from that area uses at least some of them. Linguists, on the other hand, talk of “the western Pennsylvania dialect area”, recognising that none of the local features are unique to Pittsburgh or even to western Pennsylvania. Earlier dialectologists may have been more interested in words, but present-day linguists are concerned more with accent and grammatical features, many of which are shared with other areas. None of these features is used by every local person.

             

In public discussions about the local way of talking it was the type of communication technology used that determined whether local people’s views or linguists’ views dominated. In print newspapers, where journalists decide what sources of expertise to represent, claims about Pittsburgh speech used to be attributed mainly to newcomers to the area, whose authority arises from personal experience. Increasingly, claims now come from people with various kinds of institutional authority, such as school teachers, Human Resources personnel, and linguists. People on the street are also quoted. However the headlines for newspaper articles are not usually written by the same journalist, and these often push against the voices of authority. For example, “Pitt prof finds Pittsburghese a slippy subject” juxtaposes the academic linguist with the non-linguist’s preferences for thinking in terms of local words (here, slippy) and referring to Pittsburghese.

Or the people on the street?

On a local website about Pittsburgh speech, on the other hand, anyone could act as an expert, whether or not they were from the local area. There was no way to indicate the source of a person’s expertise and no way to contest the claims of others. Although the information was presented on the web site as there for entertainment rather than as representing any linguistic expertise, this did not prevent it being drawn on by students of language and linguistics. 

                                                                                                       

Contributors to an online discussion forum, by contrast, could indicate where their expertise came from, and most of them seemed to feel the need to do so. Some told stories to show their local experience, others mentioned local linguistic features to demonstrate their knowledge and some justified their comments by referring to books or scholarly articles.

Johnstone also analysed the Wikipedia entry on Pittsburgh English, which she had developed herself. Of course, the entry had been altered and added to by other people, since this is the way that Wikipedia works. However, the volunteer editors who monitor Wikipedia rely on the citation of scholarly research as a way of evaluating entries. Although it might be thought that the Wikipedia entry would have the most potential for interaction between different kinds of expertise, in fact the voice of local people is less present there – unless they can cite the published research of sociolinguists or dialectologists ­ – than it was in the least interactive of media, the newspaper reports.  

Johnstone’s research suggests, perhaps surprisingly, that if linguists want to get their voice heard in public discussions of language, their best chance of doing so is through Wikipedia. But her research also shows that public understanding of language and dialect comes from several different kinds of expertise, of which the linguist’s can never be the only one.

 _______________________________________________________

Johnstone, Barbara (2011). Making Pittsburghese: Communication technology, expertise, and the discursive construction of a local dialect. Language and Communication 31: 3-15.

doi. 10.1016/j.langcom.2010.08.010

This summary was written by Jenny Cheshire

The numbers game

Can numbers carry social meaning?

Nikolas Coupland’s research suggests that they do. He used the Yahoo search engine to trawl through the worldwide web for numbers between one and one hundred, looking first for words (such as ninety-nine) and then for digits (such as 99). In both cases, there was a regular pattern of declining frequency of use with increasing size of numbers. Both twenty and 20 were more frequent than ninety or 90 – or even thirty or 30. Perhaps this is partly because it is easier to remember and understand small numbers, but cultural norms are also relevant. Coupland points out that we have developed ways of keeping numbers small by, for example, saying 10 minutes rather than 600 seconds, 10 pounds rather than 1000 pence, and even 4 billion rather than 4000 million. If we decide to use the larger forms, it is to make a rhetorical point.

Unexpectedly, the pattern of decreasing frequency with increasing size of the number turned out to be regular even within smaller sets of ten: 21 was more frequent than 22, 22 more frequent than 23, and so on, and this pattern was repeated within the 30s, the 40s, the 50s, 60s, 70s, 80s and 90s. What broke the pattern were the multiples of 10 (so, 30 was more frequent than 29) and, though to a lesser extent, the in-between numbers ending in 5. Other, more subtle, breaks in the pattern came from multiples of 6, numbers ending in 9, and ‘twin numbers’ like 44 or 99.

These patterns again reflect social and cultural factors. Whether or not the decimal number system originates in human biology (the number of our fingers or toes), there is now a naturalness about round numbers ending in 0 that makes them attractive when we don’t need to be precise. Numbers ending in 5 share something of the same quality. Multiples of six may also be culturally natural to us, featuring in boxes of eggs, bottles of wine or bags of bread rolls, but they have also been institutionalized as part of the measurement of angles in geometry, of longitude and latitude in navigation, and of the structuring of time into minutes and seconds.

The popularity of 99 on the web must be partly on commercial grounds, with prices such as 5.99 seeming more affordable than a price of 6.00. However this is not the only explanation. Coupland found that in the majority of cases 99 or Ninety-nine was used as a brand name. For example, a Ninety-nine card game was advertised, as was a Ninety-nine Bar and Kitchen and a book called Old Ninety-Nine’s Cave. Numbers such as 99 and 44, then, seem to have an aesthetic appeal. Even apparently neutral numbers are used more often than might be expected as brand names – seventeen, for instance, is a frequent brand name, though there seems no obvious reason why this should be more appealing than sixteen, say, or eighteen.

The study of numbers, then, raises interesting questions for students of language as well as for students of mathematics. Coupland’s study shows that it is well worth exploring the wider senses in which numbers do and do not ‘count’.

Coupland, Nikolas 2011. How frequent are numbers? Language and Communication 31: 27-37.

doi: 10.1016/j.langcom.2010.09.001

This summary was written by Jenny Cheshire

How do political leaders justify their actions?

We’ve probably all been in a situation where we felt the need to justify or ‘legitimise’ our behaviour or actions in some way. This legitimisation usually takes the form of providing arguments which can explain our actions or even our opinions about something. There can be many reasons why we do this; to gain social acceptance; perhaps an attempt to get or maintain power; to improve relationships or maybe to achieve popularity or fame. Whatever the reason, in most cases when we do this, we’re looking to get the support or approval of the person or people that we’re talking to.

A good example of where we see this process of legitimisation in action is in political discourse. Politicians (or ‘political actors’) need the support of the electorate and they must therefore justify their actions, non-actions or their ideological standing on a particular issue in order to gain that support. Researcher Antonio Reyes suggests that speakers use five key strategies of legitimisation in order to achieve their political goals. They are: 1) emotions (particularly fear), 2) a hypothetical future, 3) rationality, 4) voices of expertise and 5) altruism. To illustrate how these strategies are employed, Reyes uses examples from speeches given by two leaders in two different armed conflicts, George W. Bush on Iraq (2007) and Barack Obama on Afghanistan (2009), with both speakers seeking to justify military presence in what has become known as the ‘War on Terror’.  

Reyes says that the first strategy uses an appeal to emotions which allows the speaker to ‘skew the opinion’ of the audience about a specific issue. He states that since 2001, political leaders have often drawn on the events of 9/11 to activate people’s emotions in order to legitimise future actions. The date is so engraved in people’s memories that any mention of it sets off the emotions related to the event. So when Bush said (11 Jan 2007) ‘These are the same folks that came and killed about 3,000 of our citizens’ (all bold font appears in original article) and Obama said (1 Dec 2009) ‘On September 11 2001, 19 men hijacked four airplanes and used them to murder nearly 3,000 people’  Reyes suggests that the sentences trigger emotions which are ideal to later ‘legitimise political actions based on the effect of those emotions’.   

Another strategy that Reyes identifies is one of posing a hypothetical threat in the future to justify actions in the present. In this case the political actor usually refers to a cause in the past (here 9/11) which may happen again in the future and which therefore requires action in the present in order to stop that future threat. Reyes provides an example from Bush (11 Jan 2007) when he said ‘If we were to fail in Iraq, the enemy would follow us here to America’.

The third strategy is legitimisation through rationality. Here, the political actor is said to present the legitimisation process as a process where all the options have been considered and a decision has been reached after a ‘heeded, evaluated and thoughtful’ procedure. Linguistically, the strategy is often framed by clauses such as ‘after consultation with our allies’ or verbs relating to mental and verbal processes such as ‘explore’ and ‘consult’.

The fourth strategy is to draw on voices of expertise to strengthen the politician’s position. An example given comes from Bush (11 Jan 2007) when he said ‘and so our commanders looked at the plan and said, Mr President, it’s not going to work until – unless we support – provide more troops.......’ where a direct quote is used from authoritative figures who have been consulted for their expert opinions.

Reyes final strategy is one of altruism, meaning that political actors will often propose their actions as being beneficial to others. Examples are taken from the speech of Bush (11 Jan 2007) who said ‘It’s time to act not only for our sake, it’s time to act for the sake of the people in Iraq’ and Obama (1 Dec 2009) who said ‘Thanks to their courage, grit and perseverance, we have given Iraqis a chance to shape their future, and we are successfully leaving Iraq to its people’. Reyes states that actions are more likely to be accepted if they are framed as being beneficial to others, particularly towards the poor, innocent or vulnerable members of a society.

In his conclusion, Reyes claims that even though these two political figures have differing ideologies and that they use very different styles to present their arguments, they nevertheless employ similar strategies to legitimise their political agendas. Of course, political discourse may be said to be planned or pre-planned and Reyes calls for further studies to test the strategies in other speech events, suggesting that such strategies may also be used in casual conversation.  

_________________________________________________

Reyes, A. Strategies of legitimization in political discourse: from words to actions. Discourse and Society 22(6) 781-807.

DOI: 10.1177/0957926511419927

This summary was written by Sue Fox

Are we putting ‘People First’?

disabled person or person with disability?

Back in the early 1990s, there was a movement towards the modification of language used to describe socially sensitive issues.  This Politically Correct (PC) language was designed to reflect changing attitudes within and across societies and generate neutral terms free of negative associations.

Helena Halmari looked at a particular manifestation of this known as ‘People First’ language.  This is when pre-modified nouns are replaced in favour of post-modified nouns.  This is the difference between, for example:

1)       The disabled people: pre-modified noun.  The adjective disabled comes before the noun people (non-PC)

2)      The people with disabilities: post-modified noun.  The noun comes before the modifying prepositional phrase with disabilities (PC)

Example (1) is considered the non-PC form as the description is placed before the noun, giving it priority over what it’s describing.  On the other hand, example (2) is considered the appropriate PC form as the noun comes before description, giving the noun more status.  It is this structure of noun first which leads to the term ‘people first’.

Conducting her research in the US, Halmari looked at the use of PC and non-PC structures in the Houston Chronicle (the 7^th largest paper in the US) and internet-based Google News reports between 2002 and 2007.  She wanted to see how widely the new PC form had been adopted in the written media a decade or so after it was first introduced.

To do this, she selected two highly sensitive terms, ‘mental’ and ‘retard’ and searched both new sources to see how they were used.  To Halmari’s surprise, out of the 545 relevant phrases in the Houston Chronicle, 74% used the discouraged, non-PC structure (e.g. Mentally ill inmate’s execution date nears).  However, the use of these forms was not random.  The non-PC forms were heavily favoured when the descriptions were of fictional characters and ‘undesirable societal elements’, such as criminals.  There was also a significant link between the use of a non-PC form and the words execute and executions (once again, providing a link between this structure and criminality).

On the other hand, the data showed that the use of the PC structure (26% of the data) was higher when those being described were children and non-criminal adults.  The amount of PC forms was also boosted by their use in the names of organisations which provide a social service (such as a day program for adults with disabilities). These can be said to give the data a false interpretation as they appear frequently.

Similarly, the data from Google News showed a higher percentage of non-PC forms (only 100 out of 373 examples could be classified as PC), which were mostly found in similar circumstances to those in the Houston Chronicle.

Possible reasons for news writers favouring non-PC forms may include the fact that headlines need to be direct and eye catching to snag readers’ attention, and these forms tend to use fewer words and take up less space. However, the data analysed by Halmari suggest that the new ‘people first’ language isn’t being applied to forms across the board.  Instead, there seems to be a split in its application (i.e. depending upon who is being described) which is likely to prevent it being fully adopted by the written media. 

_______________________________________________________

Halmari, Helena (2011) Political correctness, euphemism, and language change: The case of ‘people first’. Journal of Pragmatics 43: 828–840.

doi: 10.1016/jpragma2010.09.016

This summary was written by Jenny Amos

Gestures make a difference

Children learn new words more quickly when their caregivers use symbolic gestures

Researchers have known for some time that young children’s vocabulary grows more rapidly when their caregivers talk a lot to them. So far, though, there has been little evidence of a direct link between the specific words that caregivers use and the speed with which their children acquire them.

Now Maria Zammit and Graham Schafer have shown that such a link really does exist. Not only that, their research suggests that children learn a word even more rapidly when their caregivers gesture at the same time as saying the word.

Zammit and Schafer recruited 10 white middle class mothers, all with babies aged 9 months. They chose ten everyday nouns for the study: apple, book, boot, bowl, carrot, cat, cup, dog, pencil and shirt. Every month, the mothers and their children were filmed in two situations. In the first, the researchers projected pictures of five of the objects onto the wall, one by one, and asked the mothers to talk to their child about each object. After 20 seconds they asked the mothers to stop. In the second situation, the word itself was projected, rather than a picture. The next month the other 5 objects were presented in the same two ways. This procedure was repeated until the children were aged 2 years and 2 months – past the age at which children have normally acquired these everyday words. The mothers were asked to keep a record, at home, of how old their children were when they first seemed to understand each of the ten words, and how old their children were when they first uttered each of the words.

Not surprisingly, the children who learned to understand the words at the earliest ages were those whose mothers had talked longest about the object. There was a similar though weaker association between the amount of speech from the mother and the age at which their child actually produced the word, reflecting the fact that comprehension usually comes before production. But the children who both understood and produced the words at the earliest ages were those mothers who frequently labelled an object while talking about it– in other words, who frequently used the word cat when talking about a cat.   

What was surprising was that children also acquired the words at an earlier age when their mothers not only labelled the object frequently but also made symbolic gestures to describe it (such as, for pencil, miming the act of writing). Simple pointing gestures at the picture of an object on the wall were less effective.

Zammit and Schafer suggest that gesture helps children acquire a word because it captures the child’s attention, making the mother’s speech more memorable. The gesture also adds information that can help children to retain and recall the association between the word and the object that it refers to. Some mothers seemed intuitively to realise that using gesture to describe an object while they labelled it, even before their child could understand the label, would help their child to learn the word.

 ______________________________________________________

Zammit, Maria and Schafer, Graham 2011. Maternal label and gesture use affects acquisition of specific object names. Journal of Child Language 38: 201-221.

doi 10.1017/S0305000909990328

This summary was written by Jenny Cheshire

It’s a lorra lorra laughs!

Young people in Liverpool are using a lorra t-to-r compared to previous generations

Cilla Black’s well-known catchphrase of the 1970s contains the Liverpool English ‘t-to-r’ feature. This is where, in connected speech, the /t/ sound in words such as lot and get become an /r/ sound; for example: get off sounds like ‘gerroff’ and lot of sounds like ‘lorra’. This pronunciation feature has been observed in Tyneside and West Yorkshire English, as well as in Liverpool. 

Lynn Clark and Kevin Watson investigated its use in recordings of Liverpool English, noting that it is often in competition with the glottal stop (such as ge’off) as well as with /t/. However, they also noted that not every word or phrase that could undergo t-to-r does so in natural speech. Also, other research has shown that some words are more likely to undergo t-to-r than others. When this feature was investigated in West Yorkshire English by Judith Broadbent, she proposed that words which were high frequency (i.e. words which are very common in language, like but) are more likely to undergo and preserve t-to-r than those which are lower frequency (i.e. not so common in language, like set). 

Clark and Watson wanted to test this idea in their data, as well as seeing whether there was a change occurring in Liverpool English towards one particular pronunciation.  Across three locations in Liverpool, they found that the recordings of people born in the early 1900s showed significantly less t-to-r than their modern recordings of adolescents. The adolescents were extremely consistent with their application of t-to-r, even preferring it over the glottal stop in contexts where it comes between vowels and at the end of a word, such as in a phrase like a lot of. It looks, then, as though a language change has taken place in Liverpool over the course of the last century.

Looking more closely at the variability shown by the speakers in the older recordings, Clark and Watson tested to see if word frequency was playing a part in their t-to-r use.  In order to calculate how common a word is in the language, they used the British National Corpus to decide whether a word should be treated as high or low frequency.  They showed that t-to-r only occurred in words which were classified as high frequency (like bit) and not in low frequency words.  Therefore, this supported the above claim by Broadbent.  However, when they looked at only the high frequency words, they found that, even if a word was more common, it didn’t necessarily mean it had the highest rate of t-to-r.  The variability in the older recordings confirms that there was a change in progress, towards increasing use of t-to-r, but the pattern of change is more complex than a simple frequency correlation.

The researchers concluded that, while a broad frequency comparison showed a strong difference between whether a word could undergo t-to-r or not, the behaviour of individual high frequency words was variable.  Other linguistic factors, such as the grammatical class of the words and the type of vowel preceding the /t/, did not affect the extent of t-to-r.  Therefore, even though t-to-r variation is affected by how often we say and hear a particular word, we need to look more deeply into the mechanisms of our linguistic systems in order to find the answers to what exactly causes a person to say a lorra cards but not a serra cards.

__________________________________________________________

Clark, L. and Watson, K. (2011) Testing Claims in a Usage-Based Phonology with Liverpool English t-r; English Language and Linguistics 15: 523-547

doi 10.1017/S1360674311000153

Written by Jenny Amos

Um, are you telling the truth?

Finding truth among  the lies?

You might suspect that someone is not telling the truth, but is there any way that you can be certain? Most people rely on intuition or they look for changes in body behaviour, such as fidgeting or breaking eye contact, but it seems that humans are not very efficient lie detectors and even those who have to make judgments about truth and deception as part of their professional role only perform at the level of chance. Some researchers have therefore turned their attention to more quantifiable cues that distinguish truth from deception and which do not rely on the human observer.

Researchers Gina Villar, Joanne Arciuli and David Mallard analysed the use of ‘um’ in the truthful and deceptive speech of a convicted murderer which he produced in two different contexts: when he was speaking in a formal media interview and when he was speaking in a (secretly taped) telephone conversation with his mistress. There are two hypotheses with regard to the use of ‘um’. One suggests that there would be an increased use of ‘um’ during deceptive speech because it reflects increased emotional and cognitive effort while telling a lie. The other hypothesis suggests that there would be a decreased use of ‘um’ during deceptive speech because deceivers might deliberately control their use of ‘um’ in order to appear more fluent and hence more credible. The researchers had found in a previous study that in laboratory elicited lies (where the lies were considered to be low-stake) participants had revealed a significantly decreased use of ‘um’ during deception and they wanted to see whether these results would be corroborated in real life data and where the lies were more high-stake.  

The researchers analysed the transcripts of four televised media interviews (each 20-30 mins long) and around 11 hours of secretly taped telephone conversations between the convicted murderer and his mistress (who had agreed to the conversations being taped). The transcripts were then carefully read in order to isolate all the utterances that could be verified as being either truthful or deceptive. Each sample was then coded for the presence of ‘um’ which was calculated as a percentage of the total number of words per sample.

The results clearly showed that ‘um’ was used less frequently in the deceptive speech compared to the truthful speech. The result held in both production contexts of the formal media interviews and the informal telephone conversations. The results, then, were in line with the researchers’ earlier findings with regard to low-stake laboratory elicited lies. While a single case study may not be generalizable to other persons, there is nevertheless some evidence from these results to suggest that a word such as ‘um’, which is often considered to be a ‘filler’ or unplanned error in speech, may be under the strategic control of the speaker.  The findings suggest that, in an attempt to successfully deceive, people can manipulate their linguistic behaviour and that ‘um’ may have a more important role in speech than many people realise.

_____________________________________________________

Villar, G., Arciuli, J. and Mallard, D. (2011). Use of “um” in the deceptive speech of a convicted murderer. Applied Psycholinguistics. 1-13.

doi:10.1017/S0142716411000117

This summary was written by Sue Fox

Uh, more on the mysterious case of 'uh' and 'um'

A recent summary on this blog (Er, what about this?) discussed the intriguing finding that while male speakers of British English used um and uh (or erm and er, in British English) more often than female speakers, females preferred um over uh. Now recent research in the US has revealed that female speakers of American English behave in the same way – at least in the two sets of data that Eric K Acton analysed.

Acton analysed two distinct corpora of spoken American English. One was a collection of 992 audio recordings from three speed-dating sessions held for graduate students in 2005. He found small but statistically significant differences between the overall rates of um and uh in men’s and women’s speech, with men using them more frequently, just as in Britain. A far more dramatic difference, though, was in the proportion of um to uh in men and women’s speech. This was more than three times higher in the women’s speech than in the men’s speech. 

The second corpus was the Switchboard Corpus, a database of more than 2400 telephone conversations between people across the USA, recorded in 1990. Again Acton found that the proportion of um to uh was higher for women than for men. In this case, the proportion of um to uh in female speech was two and half times as high as in male speech – not as high, then, as in the Speed Dating corpus, but sizeable nonetheless. The Switchboard Corpus includes conversations from different regions of the country: although the degree of women’s preference for um over uh varied across the country (it was highest in New England and lowest in the South), the gender differences persisted across the different regions. Acton considered the possibility that the gender of the listener might affect the use of um rather than uh: however, while both men talking to men and women talking to women used higher proportions of um relative to uh than when talking to the other gender, the proportion of um to uh for men talking to men was less than half that of when women were talking to women. As in the research on British English, younger speakers in the Switchboard Corpus used more um than uh compared to older speakers, suggesting that a language change may be occurring towards the use of um rather than uh. 

Like previous researchers who have analysed um and uh, Acton is unable to find an explanation for the dramatic gender differences in his data. He notes that he now intends to investigate whether um and uh may differ in what they communicate. He does not expect there to be an explanation as direct as “um means ‘female’” and “uh means ‘male’”. This would run counter to other research on the relation between social categories such as gender and social meaning, and would not account for the different frequencies he found among speakers of different ages (nor amongst speakers of a different social status, as found in the research on British English). Perhaps there is something about the meanings of um and uh and women’s relation to society (in both the US and Britain) that can explain why women seem to leading a change towards the increased use of um rather than uh. Even so, Acton would not expect all women or all men to behave in the same way, and exceptions to the general rule may turn out to be just as informative as the original generalization.

Acton concludes that um and uh, both in their ubiquity in spoken English and the degree to which their use is socially stratified, provide a rich site for understanding the dynamics of language use and social meaning. Certainly they present an intriguing puzzle for researchers.

Acton, Eric K. 2011. On gender differences in the distribution of um and uh. University of Pennsylvania Working Papers in Linguistics 17/2.

This summary was written by Jenny Cheshire

Er, what about this?

er and erm have important functions in speech

We may think of er and erm (or uh and uhm as they are usually represented in American English) as unimportant little fillers, but Gunnel Tottie’s research suggests that they have important functions in speech. Her research also reveals some intriguing social differences in the way that people use them.

Tottie analysed two collections of spoken English: the impromptu conversational section of the British National Corpus, and the more context governed part of the same corpus, which contains transcripts of talk from domains such as business and education. Overall there were more ers and erms in the context governed collection of recordings, but in both sets of data men used them more frequently than women. People over the age of 60 used more ers and erms than younger speakers, and so did speakers who were better educated and from a higher social class.

Even more surprisingly, although er was slightly more frequent overall than erm, erm was used more often by women than by men. The pattern was very clear in both sets of recordings. Not only that, there was a clear social class distribution, with erm accounting for just over 30 per cent of the total number of er and erm in the speech of the lowest social class but rising steadily across the different social classes to reach nearly 50 per cent for the highest class. There was a similar steady increase across different age groups, with the youngest age group using the highest proportion of erm and the oldest age group the lowest proportion. This could suggest, Tottie points out, that a language change is occurring in spoken British English, with erm gradually taking the place of er.

Why do these social differences exist? Tottie admits that her research is a preliminary study, and that more detailed analyses of the way people use er and erm are needed to answer this question. One intriguing avenue of enquiry, she suggests, could be the fact that erm tends to occur before longer pauses in speech, while er occurs before shorter pauses. The question then is why speakers pause at all, and why they utter er or erm to signal that they are pausing rather than staying silent.

Tottie points out that although these little words are often thought to mark hesitation while speakers search for the word they want to utter, they have important functions in speech for listeners as well as for speakers. Using er and erm gives speakers time to plan their utterance and shows the listener that despite the pause the speaker is intending to say something more. But at the same time er and erm help organise the utterance for the listener. In experiments, people are better able to remember a word when it has been preceded by er or erm. It seems that these ‘fillers’ prepare listeners for the introduction of a new concept. They also indicate the structure of the utterance so that listeners are better able to follow the arguments. The fact that er and erm were more frequent in the context governed collection of transcripts, which includes speeches and talk from meetings rather than impromptu conversations, probably reflects the extra effort that speakers expend on planning what they are about to say in these kinds of contexts.

Tottie’s view is that although er and erm are usually referred to as ‘fillers’, a more positive term that better reflects their function would be ‘planners’. Planning is a fundamental characteristic of intelligent behaviour, and speakers’ planning gives listeners time to figure out what will come next.

Tottie. Gunnel 2011. Uh amd uhm as sociolinguistic markers in British English. International Journal of Corpus Linguistics 16.2: 173-197.

doi: 0.1075/ijcl.16.2.02tot

This summary was written by Jenny Cheshire