So like, why do they keep saying “um” so much?

Ellee Vikram

College is a time for learning, but learning what? Surely we can attest that only a select few of us really remember college calculus. College is really a place to learn who you are and why you are the way that you are. But what shapes this? Arguably, the language and linguistic styles that you use are quite indicative of the identity you associate with, which is altogether very fluid. This study focuses on the relationship between the gender and sexuality identities of college organization student leaders and their rates of use of like, uh or um when placed in a leading speaker or a replying speaker role during a board meeting. Specifically, I am looking at the differences in the ratios of like, uh or um to all other words spoken by directors of the club Asian Pacific Health Corps at UCLA when in different speaker roles who also have different gender and sexuality identities. The study found that women and bisexual men have a higher prevalence of like, uh or um than straight men, and that there is a higher prevalence of like, uh or um when the speaker is in the leading role rather than the replying role regardless of gender or sexuality. These findings suggest that filler words serve as methods of indicating speech roles while also contributing to gender and sexuality through their purpose in defining performative identities. So, in this way, college students are shaping their identities through the way they speak and how they utilize like, uh or um.

Introduction and Background

Have you ever counted the number of times your professor said “um” during a lecture? Has it ever bothered you, or do you feel that it is a sign of poor public speaking skills? If so, that’s totally okay. In some ways, filler words, which are the likes, uhs, and ums (2, 6, 7) that seem to slip through the cracks of speech, are indeed indicative of confidence level when speaking (12,14), but they can also mean so much more. In linguistics, filler words can serve as ways of holding the stage while speaking (5), as if to let the audience know that the speaker isn’t quite done yet. In applied linguistics, filler words can serve as markers for gendered language theories (14), which seek to explain the differences in male and female speech patterns and how it ties into gender dynamics.

This study is looking at the instances of filler words spoken by college student leaders of Asian Pacific Health Corps (APHC), a pre-health organization at UCLA led by ten directors who are Asian American college students majoring in the sciences. On this board, four identify as male and six identify as female. Two females and one male identify as bisexual while the others identify as straight. At board meetings, each director is in charge of a specific section of the club, so each director has a chance to be a leading speaker (who leads a discussion) while the others assume replying speaker roles (who simply provide feedback and follow the leading speaker’s topics).


An audio recording of an APHC board meeting was taken and transcribed (1) at various time points that best highlighted the usage of filler words or clearly indicated what position the speaker was in. The instances of filler word usage (like, uh, or um) for each example chosen for the study were counted and reported as a ratio of filler words to all other words spoken for ease of comparison. These transcriptions were analyzed for linguistic features using conversation analysis (8), and were then related to larger gendered language study theories and concepts.


From linguistic analysis of the transcriptions, it became evident that females and bisexual males shared a trend in filler word usage that was different from the straight male usage when in the leading speaker role, but when in the replying speaker role, all individuals use fewer filler words overall.

Figure 1 shows that the straight female and bisexual male in the leading speaker role have a higher ratio of filler words than the straight male in the leading speaker role, while the replying speaker role shows a lower ratio overall. The deviant case is a straight male in the leading speaker role; it is an outlier due to other circumstances surrounding the example, such as supplemental linguistic features and social pressures.

The horizontal axis indicates the individual observed, and the vertical axis is the ratio of filler words to all other words.

The example in Figure 2 highlights the single deviant case that was found in this study where a straight male showed filler word usage similar to females and bisexual males. This case is noteworthy because it gives some insight into factors that can affect speakers more strongly than gender or identity performance. Here the speaker is admitting to a fault and comes off as trying to be reserved about it, and the frequency of pauses that supplement filler word usage that push his performative identity to be less dominant and less confident, which is somewhat opposite of what is expected from a straight male in the leading speaker position.

M2 = Straight Male 2

After reading the transcription it becomes apparent that the male in this case is leading conversation but is admitting to a few faults. Admitting to faults is often tough, so a linguistic tactic is to use politeness by reducing authority and assertiveness, which is something that the filler words used in this example are doing, along with the various pauses present in the speech.

Overall, the results suggest that there is some underlying difference in filler word usage between genders and sexualities depending on the speaker role, but what importance does this hold?


In gendered sociolinguistics, the Difference Model is a theory postulated by Deborah Tannen that suggests that men and women have different cultures of speech entirely (think Spanish and Mandarin, for example) which ultimately explains why men’s speech will oftentimes dominate women’s speech (15). This then bleeds into the Performativity Model, as postulated by Judith Butler, which is applicable to the actual identity that each individual is perceived as by others (3). In the case of APHC, this study suggests that college organization student leaders indeed model differences in language according to gender and sexuality, and fit into both models in some ways. This study is a very suitable example of how men and women express themselves in different ways even when in the same environment under the same pressures (9), and how strongly sexual identity can play into the gendered “culture” that one aligns with (11, 13).

Beginning with the Difference model, the results suggest that straight males use fewer filler words than females and bisexual males in the examples studied from APHC. Filler words, as mentioned briefly before, can act as markers for lessening assertiveness when speaking, which is quite common for women as women’s speech will position them in the less assertive role simply because it is their “culture” to do so, according to the Difference model (9, 15). In Figures 3 and 4, examples of straight male speech and straight female speech in the leading speaker role from the APHC board meeting are compared, and it is evident that there are multiple linguistic features at play:

M2 = Straight Male 2

As seen in transcription in Figure 3, the usage of filler words is fewer and the assertiveness of getting the point across is more evident, as this individual simply jumps into saying what they need to say without too much hedging. The assertiveness is a common feature in male linguistic culture.

F2 = Straight Female 2

In example in Figure 4, filler word usage is higher than the previous example involving the straight male speaker. Filler words here are causing the speaker to be less assertive in getting her words across, and the pauses further push the speech to seem less dominating.

Most of the filler words used by the female in the example pictured in Figure 4 are like, which in this case serve as “approximative adverbs.” These words increase the flexibility of interpretation of the idea by not confining it to any solid concept (6, 13). This is different from the male speech example where he mostly cut straight to the point without approximations, which ultimately highlights the Difference model in play, showing that males and females have different speech methods, even when in the same role, that lead to males having a more dominant character (6, 9, 15).

Lastly, the Performative model comes into play when looking at the example with the bisexual male in the leading speaker role. The transcription is shown in Figure 5:

BM1 = Bisexual Male 1

The ratio of filler words to all other words is similar to the ratio seen in the straight female example (0.11), suggesting a close relationship between these two examples from an applied linguistic standpoint.

Again, this example shows filler words accompanied by numerous pauses. Looks familiar, yes? In this case, the individual’s performative identity is more similar to the female in that there is far less assertiveness in this segment of speech, which is actually expected from this individual as bisexual and gay males tend to have performative identities comparable to females (11, 13). In this case, the linguistic features support such an idea because the same patterns of filler words and pauses are seen, and they serve the same purpose in reducing the assertiveness of the individual speaking.


In conclusion, filler words are actually quite important despite being overlooked so often. Their usage alone often is not very definitive of any purpose, but when combined with other linguistic features and thought of from an applied sociolinguistic standpoint, they suddenly become appreciably important in describing gender and sexuality identities of individuals through language.



  1. Atkinson, J. M. & Heritage, J. (1984). Transcript notation. In J. M. Atkinson & J. Heritage (Eds.), Structures of social action (pp. ix – xvi). CUP
  2. Bortfield, H. et al. (2001). Disfluency Rates in Conversation: Effects of Age, Relationship, Topic, Role, and Gender. Language and Speech, 434 (23), 123–147.
  3. Butler, J. P. (1993). Bodies That Matter: On the Discursive Limits of Sex. Routledge.
  4. Christodoulides, G. (2016). Effects of Cognitive Load on Speech Production and Perception. PhD Thesis, Université catholique de Louvain.
  5. Duvall, E. D., Robbins, A. S., Graham, T. R., & Divett, S. (2014). Exploring filler words and their impact. Schwa. Language & Linguistics, 11, 35-49.
  6. Eckert, Penelope, and Sally McConnell-Ginet. (2013). Language and Gender. Cambridge University Press.
  7. Fuller, J. M. (2003) The influence of speaker roles on discourse marker use. Journal of Pragmatics, 35, 23–45.
  8. Hoey, E. M. & Kendrick, K. H. (2017). Conversation Analysis. Research Methods in Psycholinguistics and the Neurobiology of Language, 152-189.
  9. Koike, D. A. (1986). Differences and Similarities in Mens and Womens Directives in Carioca Brazilian Portuguese. Hispania, 69(2), 387. doi: 10.2307/341699
  10. Lakoff, R. (1975) Language and Women’s Place. New York: Harper and Row.
  11. Pennington, S. (2009). Bisexuals “Doing Gender” in Romantic Relationships. Journal of Bisexuality, 9(1), 33–69. doi: 10.1080/15299710802660029
  12. Siegel, M. E. A. (2002). Like: The Discourse Particle and Semantics. Journal of Semantics, 19(1), 35–71. doi: 10.1093/jos/19.1.35
  13. Smyth, R., Jacobs, G., & Rogers, H. (2003). Male voices and perceived sexual orientation: An experimental and theoretical approach. Language in Society, 32(3), 329–350. doi: 10.1017/s0047404503323024
  14. Shriberg, E. (1996). Disfluencies In Switchboard. Proceedings of the International Conference on Spoken Language Processing.
  15. Tannen, D. (1990). You just don’t understand: Women and men in conversation (p. 42). New York: Morrow.

Read more

Leave a Reply

Your email address will not be published. Required fields are marked *