What Speakers Don’t Notice: Language Changes Can Sneak In

Leslie Barratt (Indiana State University, USA)
[BIO]

This paper addresses a recent language change in U.S English. The surprising aspect of this change is that it went unnoticed while it occurred among younger speakers in different parts of the country while older speakers continued to use the older form. The change was one affecting replacement in only a small part of the lexicon: that of by accident to on accident. The paper will also show that while by accident is still more common overall than on accident, and while older speakers still use by accident and often do not accept on accident, among younger speakers of U.S. English, on accident is common and, in fact, has equal acceptance with by accident.

1. Introduction

It is commonly acknowledged that variation in the use of prepositions among regional and social varieties of U.S. English is quite common (e.g. different than versus different to versus different from, quarter to versus quarter till versus quarter of, on line versus in line, to name a few), yet Wolfram & Schilling-Estes (1998, 61 and 343-44) have noted that variation among prepositions is not generally considered socially significant. Such is the case with the variation of the innovative on accident and the corresponding older form by accident.

Pedagogical grammars have traditionally taught students that the only preposition to use before accident is by: I spilled the milk by accident. Likewise, some dictionaries (the OEDO, for example) cite examples of by with accident. But now we also hear on accident. How did the form on accident take hold in such disparate areas in the U.S. as Terre Haute, Indiana, Farmington Hills, Michigan, Irvine, California, and McRae, Georgia? These four communities are not just geographically far apart; they are also quite different in terms of size, economics, social classes, and ethnic diversity. Yet, this paper will report on a study which demonstrates that the same change from by accident to on accident took place within the same age groups during the same period of time in different parts of the country and was unnoticed as it occurred among younger speakers while older speakers continued to use the older form.

I began studying on accident about ten years ago after I heard it separately from both my daughter, who was growing up in Indiana, a Midwestern state in the U.S., and her cousin, who was growing up in California on the west coast. While speaking to many local middle schoolers about language variation, I informally surveyed them and found that the majority used on instead of by, so I had my sociolinguistics class at Indiana State University help construct and administer a survey of local speakers in Indiana. Then, in order to see whether younger speakers using on accident is a regional or national phenomenon, I later conducted the same survey among at least younger speakers in three other U.S. states, namely Michigan (on the northern border of the U.S. but still in the eastern part of the Midwest), California (on the west coast), and Georgia (on the east coast).(1) TABLE 1 shows the socioregional data for all four states.

2. General Methodology

The survey (see FIGURE 1) was designed to elicit actual usage, reported usage (what respondents reported they would say) and reported acceptance of on accident and by accident.

In all four states, most participants took only a few minutes to complete the survey. Participants were first asked to fill in the missing words in Part I, which includes two close items intended to determine actual by/on accident usage: "I did it ___ accident" and "I opened it ___ accident." Participants then turned the paper over and answered the questions in Part II, which provides information on reported usage (#1 "If someone breaks your window, and it’s not purposely, it happened (A) by or (B) on accident") and reported acceptance: #2 "Is the other choice in #1 one you would ever use?"

3. The Indiana Survey

My students received completed surveys from 179 mostly working and lower middle class subjects in Terre Haute, Indiana, an industrial town of approximately 60,000 inhabitants in one of the ten poorest counties of the 92 in the state. Subjects included middle school, high school, and university students, as well as employees in a local factory and residents of a retirement center. The data analysis was conducted using SPSS 10.0 and yielded positively correlated results between age and linguistic preference. The results for I did it ..... accident are shown in TABLE 2.

These results can be seen more easily in the line graph in FIGURE 2.

The age ranges on this graph represent 5 years, so 1 represents 6-10 year olds, and 5 represents 26-30 year olds. Results for the second cloze item are shown in TABLE 3 and FIGURE 3.(3)

FIGURE 3
Indiana Result for I opened it ..... accident.

4. Use

As we can see, the distribution of on accident is associated with age such that usage of both on and by occurs by people under the age of 30 (age range 5 and below), but usage of on is almost non-existent among people older than 30 (6 and above). This difference is significant at the .003 level (Chi-Square) for I did it ..... accident and at the .0000 level for I opened it ..... accident. In addition, 17 subjects (9.5%) used by in one answer and on in the other. Of these 17 subjects who used both, 15 were aged 25 or under; the other two were 56 and 62, making the average age for these speakers 23.4 years.

5. Reported Use

The results from Indiana are quite clear: on has emerged as a common preposition before accident among speakers under the age of 30. As with any language change, however, there is a question of what speakers use and what they think they use. Part II of the survey asked subjects specifically, "If someone breaks a window, and it's not purposely, it happened (A) by accident or (B) on accident."

The results (in TABLE 4 and FIGURE 4) show that again, younger speakers report using on and by while older speakers report almost exclusively by. The results for Parts I and II are not identical, however. In fact, 19 of the 65 people who used only on in Part I (29.2%) reported that they would use by in Part II while only two people who used only by reported in Part II that they would use on. Of the 17 people who used both, 15 reported they would use both, but two 13-year olds reported they would only use by. Clearly, some speakers are not aware of the form that they in fact use.

6. Acceptance

The last part of Part II asked speakers if they would ever use the other preposition: "Is the other choice in #1 one you would ever use?" TABLE 5 collapses answers A and B (i.e. yes and possibly and/or in some situations) to contrast these with C (no), which is a non-acceptance of the other preposition.

The numbers of respondents accepting the other choice as at least possible in their speech were relatively high for respondents who exclusively used by and for those who exclusively used on, but we can see that users of on are much more accepting of by than by users are of on (91% versus 46%, and speakers of both fall in the middle (88%)). The average age for people accepting the other choice varies also: It is 37.4 for by speakers but only 15.2 for on speakers and, as mentioned above, 23.4 for speakers who used both.

Obviously, of course, this greater acceptance of the other form by on users may be related to the fact that they are generally younger, and looked at from the perspective of age, subjects 25 years old and younger are more accepting of both forms since even though by is still more commonly reported than on, 91 % of respondents 25 and under answered that they could use the other preposition. We can see from the results so far that on accident is not only in common usage in Indiana, it is in reported usage and reported acceptance as well. The surveys in Michigan, California, and Georgia were conducted to determine to what extent on accident also exists among young speakers in other parts of the U.S. besides Indiana. As a result, these later surveys did not aim for an even age distribution, although in California, it was possible to get large numbers of older speakers.

7. The Michigan Survey

While Michigan would not seem to be in another part of the U.S. since it is Indiana’s northern neighbor, Michigan is on the other side of the traditional Midland/Northern dialect boundary, and it is considered by Labov and others to be in the Inland North, the northern tier of states, whose speakers behave differently from those in other states with regard to many features, including values, attitudes, politics, and language. To provide yet more contrast with the Indiana population, I surveyed a community at the opposite end in terms of social class - Farmington Hills, Michigan, a growing town of approximately 80,000 people just north of Detroit in Oakland County, the third most prosperous county in the U.S. 197 middle school students answered the first cloze answer of Part I, I did it ..... accident, as given in FIGURE 5 by actual age rather than age groups.

Figure 5

Obviously, both by and on are quite common among the 10-14 year olds for I did it ___ accident, although the number of people who wrote by (125) is about twice as high as for those who used on (65). For the second cloze answer, the numbers were almost identical to the first - roughly two to one in favor of by. As in Indiana, in Michigan, reported use parallels actual use with the same two to one ratio. Likewise, the other choice in Michigan was generally accepted. Among those who reported that they used by (131), about 85% reported that they would use on while among those who reported that they used on (66) 92% reported that they use by. This difference is not significant at the .05 level, however.

8. The California Survey

The Indiana and Michigan data provide clear evidence that young people in these Midwestern states accept and use on accident in both working class and upper class communities. However, the question still remained of whether this form was in widespread use in other parts of the U.S. and whether it had affected the same age groups. Therefore, the survey was conducted in Orange County, California (Irvine and Laguna Hills, affluent communities of 140,000 and 31,000, respectively).

FIGURE 6
California Results for I did it ..... accident

Here 101 subjects responded from a private middle school, a tax consulting company, and an affluent retirement community. The results for the first cloze answer of Part I, I did it ..... accident, are shown in FIGURE 6.

The results from California are similar to those of Indiana and Michigan. While on accident occurs more frequently than by accident among the 11 and 12 year olds surveyed (22 to 13 for I did it ___ accident), it is completely absent among those surveyed over age 34. Likewise, in reported use, Californians were slightly less likely to report that they used on (21 responses) than they were to use it (26 responses).

Finally, people who reported that they used by were less likely to accept on than the reverse. In California this difference was significant at the .007 level, but the large numbers of older speakers (41 retirees out of 101 subjects) skews this result.

9. The Georgia Survey

Finally, after discussions emerged on the American Dialect Society Listserv regarding the presence of on accident in many places and its possible absence in the Deep South, surveys were sent to McRae, Georgia, a small, rural town of about only 3,000 residents in southern Georgia, located in Telfair County, which has the highest unemployment in Georgia. In McRae 65 people completed the survey, most of them students in an elementary and middle school. FIGURE 7 gives the results for Georgia for the first cloze answer of Part I, I did it ..... accident.

Clearly, on accident is present in the Deep South as well as in other parts of the United States, and it is common in the same age groups, especially among 9-16 year olds surveyed in McRae. This association with age is significant at the .004 level.

As in the other states, some adults use on accident too: the oldest Georgia speaker to use on accident was 38 years old. In reported use, the oldest person who reported on was 20, but among those under 20, on was actually reported more than by in Georgia. As in the other states, the majority of speakers in Georgia were willing to accept the other form, and on accident speakers were more accepting than by accident speakers but the difference was not statistically significant.

FIGURE 7
Georgia Results for I did it ___ accident.

10. Conclusion

The answer to the question about whether on accident is found nationally is clearly yes. It is prominent in all states surveyed. FIGURE 8 depicts the use of on accident (not, as in previous figures, on versus by) in all four surveys.

What can be seen clearly is that younger respondents (under the age of 30) have on accident in common usage in all four states. In fact, the only two speakers over the age of 40 using on accident were two females in Indiana, aged 56 and 62.

It should be noted that while the correlation of age with on accident usage, reported usage and acceptance was significant to as much as the .0000 level, no correlation with gender was found in any of the surveys. For both male and female respondents, on is more prevalent under age 10, both on and by are common between the ages of 10 and 35, and by is overwhelmingly preferred by those over 35.

FIGURE 8
All States results for on accident in I did it ..... accident.

Because the survey administrator in Georgia felt that ethnicity might affect someone’s use of on or by, these subjects were asked their ethnicity as one of the background questions. However, when ethnicity was examined in the Georgia data, both those identifying themselves as African Americans and those identifying themselves as non-African Americans used on accident or reported that they would use it.

What all of the studies show is that the preposition before accident varies among younger speakers of U.S. English and that the prepositional phrase on accident is common and has equal acceptance with by accident among younger speakers. The fact that some younger speakers report they would not use by accident shows that neither by nor on seems to carry any social significance for them. In many cases, younger speakers were even unaware of the existence of by accident.

The lack of stigma is also supported by the fact that so very few people (including linguists) have noticed the on form despite its being around for some time. This is not to claim that the form is never noticed (and never "corrected") by individuals (English teachers and other older speakers, for example). However, only two prescriptive grammar internet sites have included prohibitions against using on accident, and only one of these is still an active site: http://www.wsu.edu/~brians/errors/onaccident.html, which is the site for Brians (2003). Over the last couple of years, there has been a small amount of discussion on the American Dialect Society internet Listserv (ADS-L) about on acccident, yet even the ADS-L has had fewer than 100 postings about on accident since 1999. There is no mention of it in DARE, although J.H. Hall (personal communication, June 24, 2003) reports three notes about it in the DARE files from 1999-2002. The fact that so little has been written about it by linguists or by others and that virtually nothing has been written in the prescriptive usage literature about it has possibly allowed it to spread more easily as well as avoid becoming stigmatized.

The spread of on accident is further aided by the spread of on, as noted by Loringer (2000), who states that in the press on has taken on the roles of other prepositions to become "nearly an all-purpose preposition."

While it may seem unusual that a new form used by younger speakers can go unnoticed, similar phenomena have occurred as, for example, in the ‘cot/caught merger,’ which Gordon (2002) claims is neither stigmatized nor noticed.

In the ADS-L discussion, Carson suggests that some on accident users have a semantic difference between on accident and by accident with on accident meaning "accidentally on purpose" i.e. purposeful but disguised as accidental. In the context of my survey, subjects were given one meaning, yet many of my subjects used both prepositions and the vast majority of those under 25 said they could use both prepositions, so a semantic distinction such as Carson suggests was not supported by the present study.

Of course, when on accident first appeared is not answered in the present study, but there are several indications that the form is at least 25 years old (i.e. dates back to at least the late 1970s). First, there is the fact that in all states surveyed those under 30 accept and use it. Furthermore, several contributors to the ADS-L state that they first heard it by the 1980s in several parts of the U.S. (Kuhlman even recalling its use in the 1950s in a rural area outside Chicago, Illinois).

Where this change started is not clear either since on accident varies with by accident for respondents in their twenties and thirties in all states surveyed. Several scholars on the ADS-L have expressed their own hypotheses of where on accident started geographically, but these are based strictly on their having heard it first in that particular area. I should note here that the ADS-L contains reports of on accident in 17 other states: Alabama, Arizona, Colorado, Illinois, Maryland, Minnesota, Mississippi, New Jersey, New Mexico, New York, Ohio, Oregon, Pennsylvania, Texas, Vermont, Virginia, and Washington.

Obviously, wherever the form started, it was already used in different parts of the U.S. by the late 1970s, but it was not common everywhere at that time since as people moved into new areas, they heard it for the first time.

Finally, why on accident arose is also unclear. Obviously, on purpose may have played a role in supplying an analogical form (I didn’t break the window on purpose; I broke it on accident). But by accident and on purpose have existed for hundreds of years without one causing the other to change prepositions, and we don’t hear by purpose, so why did the change happen when it did and why did the change have the direction it did rather than the other way round (in other words, to by accident & by purpose)? D. Stampe (personal correspondence, August 12, 1999). also points out that some young speakers of on accident don’t use on with purpose, so analogy cannot be a clear-cut explanation.

The answers to these questions of when, where, and why cannot be answered in the present study. Simply stated, the reasons for particular language changes are rarely one dimensional, and this is no exception. Although it happened right before our eyes, it seems to have merely happened ‘on accident.’

© Leslie Barratt (Indiana State University, USA)x

NOTES

(1) I would like to thank the following people for their assistance in data collection: In Indiana: Fahad Al-Qurashi, Yuko Azuma, Kristen Bocock, Keppy Boone, Cynthia Canteen, Cheryl Cunningham, Shane Goodale, Kim Halwes, Janean Johnson, Jessica Midraj, Kristen Perry, Cindy Quackenbush-Geswein, and especially Karen Seremet, who also gathered all of the Michigan data and Brock Roberts, who also entered the initial Indiana data; in California: Cecily Burke, Eleanor Burke, Ava Taussig, and David Taussig; in Georgia: Kay Graham and Betty Phillips, who also patiently read through an early draft of this paper and gave me several helpful suggestions.

(2) We did not ask Do you accept the other choice? or "Would it bother you if someone said [the other choice]?" because such wording would have been unclear as to whether it was indicating acceptance of the preposition or acceptance of the excuse.

(3) Only the Indiana results are presented in both TABLE and FIGURE formats; data from the other states are given once in FIGURE format. Furthermore, since the results for the usage (both cloze answers), reported usage, and reported acceptance are all virtually identical, only the first usage answer is presented for Michigan, California, and Georgia.

REFERENCES

American Dialect Society Listserv (ADS-L) at listserv@listserv.uga.edu. April 1, 1997 - December 1, 2005. Archives available at: http://www.americandialect.org/index.php/amerdial/1992_1999_emaillist_archive/ (1992-1999) and http://listserv.linguistlist.org/cgi-bin/wa?A0=ADS-L (from April 22, 1999).

Barnhart, D.K. (2000, January 16) Online posting "Re: "on accident" posting #3." ADS-L archived at http://listserv.linguistlist.org/cgi-bin/wa?A2=ind0001C&L=ADS-L&P=R1297&I=-3

http://listserv.linguistlist.org/cgi-bin/wa?A2= ind0001c&L=ads-l&D=1&F=&S=&P=1473.

Brians, Paul. (2003). Common Errors in English Usage. Wilsonville, Oregon: William James & Company and at http://www.wsu.edu /~brians/errors/onaccident.html.

Carson, J. (2000, January 14) Online posting "Re: on accident." ADS-L archived at http://listserv.linguistlist.org/cgi-bin/wa?A2=ind0001c&L=ads-l&D= 1&F=&S=&P=86.

DARE. Dictionary of American Regional English. (1985-). Vol. 1 (A-C), ed. Frederic G. Cassidy. Vols. 2 (D-H)) and 3 (I-O), ed. Frederic G. Cassidy and Joan Houston Hall. Cambridge, Mass: Belknap Press of Harvard University Press.

Gordon, Matthew. (2000). Tales of the Northern Cities. American Speech, 75(4), 412-414.

Kuhlman, P. (2000, January, 18) Online posting "Re: on accident." ADS-L archived at http://listserv.linguistlist.org/cgi-bin/wa?A2=ind0001C&L=ADS-L&P=R2805&I=-3.

Labov, William. (1996). The Organization of Dialect Diversity in North America. Revision of presentation at ICSLP4, the Fourth International Conference on Spoken Language Processing at Philadelphia, October 6, 1996. available at http://www.ling.upenn.edu/phono_atlas/ICSLP4.html.

Lorringer, Paul W. (2000) The Penguin Dictionary of American English Usage and Style. New York City: Penguin Books, Ltd.

OEDO. Oxford English dictionary Online. (2002). In Oxford English Dictionary (Web site). Available from http://www.oed.com.

Wolfram, Walt, and Natalie Schilling-Estes. (1998). American English. Malden, Massachusetts: Blackwell Publishers.

1.4. Reproduktionen und Innovationen in Sprache und Kommunikation verschiedener Sprachkulturen / Reproduction and Innovation in Language and Communication in different Language Cultures

Sektionsgruppen | Section Groups | Groupes de sections

       Inhalt | Table of Contents | Contenu  16 Nr.

Webmeister: Peter R. Horn     last change: 26.6.2006