Thursday Thinkpiece: Gotthelf On, Um, Lawyers’ Verbal Fillers

Each Thursday we present a significant excerpt, usually from a recently published book or journal article. In every case the proper permissions have been obtained. If you are a publisher who would like to participate in this feature, please let us know via the site’s contact form.

The Lawyer’s Guide to Um

Barbara Gotthelf
Legal Communication &Rhetoric: JALWD Fall 2014/Volume 11

Excerpt: Introduction and Parts I through III

In 2006, my son came home from his first day of eighth grade and reported that his Language Arts teacher, whom I will call Mr. Sweeney, had made a dire announcement: He planned to charge students a nickel each time they used uh or um in class. I was pleased that Mr. Sweeney wanted his students to become better speakers, but I questioned his methods. Given my son’s pained reaction, I suspected it was going to be a quiet year in Language Arts.

Intrigued, I sent an e-mail describing Mr. Sweeney’s strategy to Elizabeth Shriberg, a psycholinguist at SRI International in California. She quickly responded and said that using uh and um was not only “perfectly normal,” but also helpful in furthering effective communication.1 As for Mr. Sweeney, she said, “Unless your language arts teacher wants to have people speak only when they’ve completely ‘rehearsed’ what to say (as in a play), he will need to allow the children to pause within their turns.”2 And if they pause, she said, the appropriate thing to do is to fill that pause with uh or um.3

Having had my maternal instincts confirmed by a psycholinguist, I filed away Dr. Shriberg’s e-mail, and the issue of uh and um receded to the background where, as it turns out, it rightfully belongs. But some years later, uh and um reappeared. In 2012, while judging a moot court competition, I heard a seasoned legal writing professor suggest to a student advocate that she had used perhaps a few too many uhs and ums. The blushing student apologized profusely. Recalling Mr. Sweeney’s methods and Dr. Shriberg’s e-mail, I silently questioned whether the professor’s criticism was useful. I began consulting books on public speaking, including texts written specifically for lawyers, and they all gave the impression that using uh and um might be the single worst thing any speaker could do. This advice seemed to be at odds with Dr. Shriberg’s email. Digging further, I discovered a body of scientific literature that supports Dr. Shriberg’s views and demonstrates that, contrary to public perception, uh and um are not only inevitable, but actually useful bits of communication.

Here, then, is my formal response—albeit delayed—to Mr. Sweeney, and a plea to speakers and listeners everywhere to give up the um fixation and focus again on the substance of our spoken communication.

I. Disfluency

A. Definition

Read a page from any deposition or trial transcript and your eyes will tell you what your ears may not: spoken English is vastly imperfect. Not only do our sentences lack the thoughtful syntax and grammar of our written work, but our speech is rife with what linguists call “discourse markers” and “disfluencies.” Discourse markers, which are viewed as actual words with distinct meanings and include like, well, you know, oh, now, I mean, mind you, everything, sort of, kind of, and so, are used in different contexts and do not appear to be interchangeable.4 Oh, for example, alerts listeners that the speaker has remembered specific information. Well indicates that a seemingly irrelevant interpretation is actually relevant, and the despised like suggests that the speaker is deliberately using vague or informal language.5 Disfluencies, in contrast, are not distinct words, but instead consist of hesitations, repeated words or phrases, false starts and restarts, and the use of uh and um. Many linguists refer to uh and um specifically as “verbal fillers” or “filled pauses.” These verbal fillers typically operate under the radar and may be missed or, more often, excised by the conscientious court reporter. The full range of disfluencies can account for up to six percent of what we utter, and the verbal fillers uh and um make up a third to more than half of all disfluencies.6

While they keep company with other disfluencies, the verbal fillers uh and um march to their own linguistic drumbeat. For example, while most disfluencies consistently increase with the speaker’s anxiety, uhs and ums generally do not.7 Longer sentences contain more disfluencies overall, but uh and um remain relatively constant.8 And, while men use uh and um more often than women, both genders demonstrate similar levels of other disfluencies. 9

These differences in usage have led linguists to study the verbal fillers uh and um as unique phenomena. For lawyers, understanding verbal fillers provides a lesson in how to approach these pesky utterances, which for most of us have come with a lifetime of stern admonishments to avoid using them at all costs. On a broader scale, the study of verbal fillers provides a glimpse into a larger world of actual spoken communication. For years, linguists resisted this world, choosing instead to define disfluencies as “errors” that needed to be eliminated from the study of pure language.10 More recently, linguists have acknowledged—and at times even embraced—disfluency as an integral part of how we communicate.

Starting with uh and um, lawyers can gain a better appreciation of what disfluency means for oral advocacy, and whether, or how, we need to address the messy reality of our spoken language.

 B. The Ubiquitous Uh and Um

Ironically, while lawyers are expected to be eloquent and well spoken, they also match the profile of the frequent ummer. Well-educated and conscious of their speech, lawyers are faced with the cognitive and social demands that almost guarantee the appearance of uh and um in their spoken communication.

The verbal fillers uh and um are defined as “verbal interruptions that do not relate to the proposition of the main message.”11 As described by psychologist Herbert Clark, speech proceeds along two parallel paths. The primary track is the “official business, or topics of discourse”—the substance of what we want to say.12 A secondary, or collateral, track refers to the act of speaking itself: “to timing, delays, re-phrasings, mistakes, repairs, intentions to speak, and the like.”13 Uh and um move along this collateral track, allowing speakers “in effect, to manage [their] on-going performance.”14

Verbal fillers appear in all languages and are typically monosyllabic with a schwa core vowel sound.15 Speakers of English use uh and um. The British spell these fillers er and um, but pronounce them the same as in North American English. Germans use äh and ähm; the French use eu, euh, em, eh, and oh, and Spanish-speaking people use eh, em, este, and pues.16 There is even a sign for um in American Sign Language.17

Virtually everyone uses verbal fillers, though the frequency can vary greatly from person to person.18 A study of one language database showed that speakers produced between 1.2 and 88.5 uhs and ums for every thousand words, with a median filler rate of 17.3 per thousand words.19 Other databases show anywhere from three to twenty uhs and ums for every thousand words, placing uh and um thirty-first in a ranking of most commonly used utterances, just ahead of or and just after not.20 A British study showed that, contrary to popular expectations, the use of verbal fillers does not indicate a lack of education or manners; instead, the use of uh and um increases with education and socioeconomic status, a finding with particular implications for the legal profession.21 Older people use more uhs and ums than younger people, and, curiously, men consistently use verbal fillers more often than women—a finding that has been replicated across several studies.22 Women, for their part, appear to use a higher ratio of ums to uhs than their male counterparts.23

For those who believe that they have eliminated uh and um from their speech—and many people hold this view—studies show that people are notoriously unable to accurately assess who is saying uh or um, or how often these fillers are being used.24 The question, then, isn’t whether we use uh and um, but why we use them.

II. The Forensics of Verbal Fillers

Broadly stated, speakers tend to use the verbal fillers uh and um when something has interrupted the enormously complicated task of speech production. That interruption can come from the difficulty of the subject matter that is being discussed, or it can come from the speaker’s selfconsciousness about the act of speaking. Researchers refer to these forms of interruption as “task complexity” and “task concern,” and both are highly relevant for lawyers, who are required to speak about difficult and abstract concepts in a stressful, and often very public, setting.

 A. Task Complexity

The theory of task complexity proposes that the use of uh and um increases along with the complexity of the subject matter the speaker is addressing. The more challenging and varied the options are for the speaker, the more verbal fillers he will use.

Linguists adhere to the widely held view that speakers use verbal fillers when they are, in effect, searching their brains for information, essentially in the same manner that a computer scans a hard drive for data.

The speaker may be looking for the next word, phrase, or idea, or making a decision about the next thought.25 Accordingly, the rate of uhs and ums increases when the topic is more abstract, as well as when the speaker is choosing from a larger vocabulary.26 As the range of options increases, so does the task complexity and the likelihood that the speaker will fill the delay imposed by the process with uh or um. This was the conclusion reached in an oft-cited study at Columbia University, where researchers counted the number of verbal fillers used by professors during lectures given to undergraduate students in three separate academic divisions: the natural sciences, the social sciences, and the humanities.27 The natural-sciences professors used the fewest verbal fillers, with a mean rate of 1.39 uhs per minute.28 The social-science professors had a mean rate of 3.84uhs per minute, and the rate for humanities professors was 6.46 uhs per minute. Yet these same professors, when interviewed on the same topic (one unrelated to their disciplines), used about the same number of fillers per minute.29 Thus it appeared that the academic disciplines, and not the speakers, drove the frequency of fillers.

The Columbia researchers concluded that lectures about natural sciences produced fewer fillers simply because the speakers had fewer options to choose from in deciding what to say. As the authors explained, there are no synonyms for molecule or atom or ion. But humanities professors—much like lawyers—have many alternatives for the words found in their discipline. A humanities professor may discuss abstract and subjective concepts like affection, class structure, prejudice, beauty, or style.30 Lawyers may discuss concepts like duty, consideration, mutuality, scienter, or malice. And, whereas concepts like E=mc2 are fixed, the options for interpreting a passage from King Lear (or Roe v. Wade) seem limitless.31 Such options increase the complexity of the task and, concurrently, the rate of verbal fillers.32

Researchers have tested the task-complexity theory in other ways, with similar results. Study subjects asked to “talk” their way through mazes used more verbal fillers when confronted with mazes that could be navigated using multiple routes.33 Conversely, mazes with a single path (and fewer choices) produced fewer fillers.34 But the maze study produced another interesting result: When study subjects were told they could use only four words to talk their way through the maze (left, right, up, down), they began to use more verbal fillers, even when describing simple mazes.35 Researchers posited that the “lexical suppression” created by limiting speakers to four words triggered a stopping and starting of the speech apparatus that prevented speakers from developing a normal speech rhythm.36 Thus, while verbal fillers are a mark of task complexity, they also appear where, “for some other reason, the flow of speech is disrupted.”37 That “other reason” may be the speaker’s self-consciousness about the act of speaking itself, or “task concern.”

B. Task Concern

While the theory of task complexity attributes the use of uh and um to the difficulty of the topic being discussed, the theory of task concern proposes that people say uh and um when something has shifted their attention away from what they are saying and moved it toward how they are saying it. In other words, the task-concern theory holds that uh and um are not so much a product of how difficult the subject matter is, but an indication of how preoccupied the speaker is with how he sounds to his audience.

The theory of task concern developed when researchers looked at word repairs, where, for example, a speaker might say, “Today is Mon–uh–Tuesday.” Researchers theorized that when a speaker detects a word error, the normal flow of speech is interrupted and the speaker becomes focused, if just for an instant, on the act of speaking.38 In that moment, the speaker produces an uh or um. This finding raised the question of whether conscious attention to the act of speaking on a broader scale, not just at the word-repair level, might make speakers use more uhs and ums throughout their speech. Two psychologists at the University of California, San Diego—Nicholas Christenfeld and Beth Creager—took a novel approach to answering this question by studying a topic of relevance to all speakers, including lawyers: anxiety.39 The researchers found that anxiety makes speakers use more uhs and ums only when the anxiety shifts the speaker’s focus to the act of speaking.

Christenfeld and Creager began by examining the popular perception that people produce more uhs and ums when they are generally anxious or nervous. If that perception is true, this situation would create a dilemma for public speakers, including lawyers, who believe that the only way to improve their performance is by finding a magic cure for their jitters. For a time, the literature on anxiety and verbal fillers was hopelessly mixed, with some studies showing that anxiety does increase filler rates, one study showing that it does not, and many studies showing no effect at all.40 Christenfeld and Creager suggested that researchers had missed the mark by failing to distinguish between the sources and the effects of anxiety. Their hypothesis was that anxiety will increase verbal filler rates only if it interrupts the automatic flow of speech by making speakers self-conscious about how they sound.41 In other words, thinking about speaking might make people “choke,” in much the same way that any conscious attention to a skilled performance can disrupt, and limit, that performance.42 Choking, in turn, would make people use more uhs and ums. Christenfeld and Creager’s hypothesis also predicted that a more generalized anxiety that is unrelated to the act of speaking would not increase filler rates.43

To test their hypothesis, Christenfeld and Creager devised a number of experiments that manipulated anxiety and self-consciousness. They consistently found that generalized anxiety did not increase filler rates, but anxiety that created self-consciousness about speaking did. In one experiment, speakers who were told that their recorded speech would be evaluated to see how creative they were—a condition designed to increase self-consciousness—produced significantly more verbal fillers than speakers who were told that their speech would be used only for routine grammatical coding.44 Another experiment compared filler rates between speakers who wore headphones that amplified their own voices and speakers who were asked to dance by themselves in front of a camera while talking.45 This study was designed to make members of the headphone-wearing group self-conscious about their speech, while making people in the dancing group anxious about an act unrelated to speech. The results showed that the speech-conscious group used more than twice the number of fillers per minute (5.61 per minute) as the anxious dancers (2.07 per minute).46 Moreover, the speech-conscious group also used more fillers than a third group that had been asked to speak on a more complex topic (3.85 per minute), reinforcing the notion that self-consciousness—or “task concern”—may create even more fillers than task complexity.47

Christenfeld and Creager confirmed their conclusions about task concern and task complexity by measuring verbal filler rates in a unique set of study subjects: people who have been drinking. Interviewing patrons at eight bars in the San Diego area, the authors accurately predicted that as people become more intoxicated, they use fewer verbal fillers.48 This observation ran counter to the notion that uh and um are markers of careless or lax speech. To the contrary, the more awareness people had of their speech, the more fillers they used. As levels of intoxication increased and people became less concerned with how they sounded, their filler rates dropped, even as the act of speaking became more difficult. In other words, task concern—the province of the sober— rather than task complexity determined filler rates.49

This work convincingly limits the idea that garden-variety anxiety increases filler rates, while also proposing a theory of ums that “moves away from the cognitive complexity notions” advanced by studies like the one done at Columbia.50 According to Christenfeld and Creager, “rather than indicating when a tough decision is being made by the normal speech production apparatus, ums may indicate when the speaker changes modes and gives deliberate attention to some aspect of the speech.”51 That deliberate attention, or self-consciousness, removes the act of speaking from its “fluid automatic mode” and triggers the use of verbal fillers.52

Christenfeld and Creager acknowledge that the theory of task concern is not incompatible with the theory of task complexity.53 A difficult cognitive task—finding the right word, thought, or idea—can make the speaker more self-conscious about the act of speech production. But they maintain that a speaker’s use of verbal fillers is more likely related to the “social context of the utterance” than to the complexity of the task.54 For this reason, Christenfeld and Creager believe that the nature of the speaker’s audience should have a profound effect on filler rates.55 Consequently, “[t]elling a story to one’s best friend may lead to fewer filled pauses than telling the same story to a parole officer.”56

III. From Symptoms to Signals

Although the production of verbal fillers may not be deliberate, research has shown that fillers may serve distinct communicative functions. Lawyers who speak before courts, clients, and other discerning audiences should know how fillers function to communicate information; they should understand that the actual effects of fillers on listeners may be less dire than imagined and may even be beneficial under some circumstances.

For many years, linguists adhered to the view that verbal fillers were merely symptoms of a breakdown in the speech process. They viewed uh and um as “errors” that fell outside the proper study of language.57 Accordingly, linguistic study was focused exclusively on the “fluent, idealized utterances” that form “an uninterrupted sequence of words that follows the rules of English syntax.” 58 But as any lawyer knows, even the most polished advocate rarely speaks in perfect prose.

Gradually, linguists began to consider that verbal fillers might not be symptoms of a problem, but actual signals used by speakers to communicate information. What were once “errors” became “verbal fillers” or “filled pauses.” Some linguists rejected even these terms as misnomers. A pause is silent by definition, making the term “filled pause” anomalous.59 As described by psycholinguists Daniel C. O’Connell and Sabine Kowal, “fillers are neither pauses nor are they used necessarily where there would otherwise be a silence; they are not a sort of putty used to fill the cracks in window frames—to stuff something into a silence. They are simply legitimate hesitations.”60 Linguists who rejected “filler” as an “uninformative default term” instead began to describe uh and um as “planners”61 or “speech management phenomena.”62 Serious study of verbal fillers produced a wealth of scientific data, and while linguists disagree on the details, a consensus emerged in which uh and um are viewed as signals that perform communicative functions.

 A. Defining Filler Functions

The study of verbal fillers began with the threshold question of whether listeners even hear them. The good news is, unless they are specifically focused on a speaker’s use of verbal fillers, listeners—including judges—will naturally ignore most uhs and ums, at least on a conscious level.63 As explained by Susan E. Brennan, a cognitive scientist at the State University of New York at Stony Brook, and her colleague Michael Schober, “much of the time, listeners don’t experience disfluencies as disruptive, and when they do detect disfluencies, they have trouble categorizing or locating them precisely.”64 In fact, listeners are notoriously unable to estimate the number or location of verbal fillers in a spoken message.65 Nicholas Christenfeld found that when study subjects listened to speech both with and without verbal fillers, their estimates of the frequency of uhs and ums were “profoundly skewed.”66 Not only did the subjects overestimate the frequency of fillers in speech where fillers actually occurred, but they guessed that they had heard an average of 22.1 ums during a three-minute tape that did not contain a single filler.67

While listeners may not be aware of most verbal fillers on a conscious level, there is strong evidence that fillers are not being filtered out to create a “sanitized,” or fluent, version of the message.68 Instead, verbal fillers are processed by listeners and used as information.69 Studies of verbal fillers demonstrate that they perform certain defined functions, which can, and often will, overlap.70 Filler functions can be divided into five categories:

1. Signal of delay: At their core, uh and um are used to signal delay, and this is their chief use as well as their “stock dictionary characterization.” 71 By using a verbal filler, the speaker is telling the listener that, for however brief a time, “I am unable to proceed.”72 The delay may be caused by task complexity or task concern, which in turn can have a number of underlying causes, as discussed above.73
   

2. Conversational signposting: Verbal fillers can be seen as signposts for people engaged in the complex give and take of conversation, which may be particularly relevant for lawyers in the context of oral argument.74 Here, uh and um can serve multiple, sometimes contradictory, purposes. Speakers use uh and um for turn-taking—to indicate that they are taking their turn to speak; for turn-holding —to indicate that they are not finished speaking and wish to hold the floor; and for turn-yielding—to give up the floor.75 Speakers help listeners distinguish between these signals by the manner in which they vocalize uh and um. For example, an uh or um that trails off can signal that the speaker has finished speaking. 76 Uh and um spoken with a rising intonation suggest that the speaker has run into trouble and is looking for help from the listener to complete a thought.77
   

3. Attracting attention: At perhaps their simplest level, uh and um can be used to attract attention to the speaker and to establish contact, as in “Uh, hello?”78

4. Highlighting: Speakers can use uh and um to focus the listener’s attention on whatever comes after the filler.79 In this sense, uh and um are “a sort of verbal italics” or “semantic booster.”80 A verbal filler that precedes a word or phrase “highlights the following element, suggests that it is being chosen circumspectly and focuses the listener’s attention on it.”81
   

5. Correction: Verbal fillers can be used to signal that the speaker has gotten off on the wrong track, perhaps by choosing the wrong word or phrase or by mispronouncing a word. Here, the speaker is indicating that “a more correct or suitable word or phrase than the one(s) just said will follow.”82 On a broader level, using uh or um can signal a change, or correction, in the topic being addressed. Either way, verbal fillers signal that the speaker intends to revise the message, and the listener should take note of the change.83 As University of Edinburgh researcher Martin Corley explains, uh and um in this context tell the listener, “‘pay attention, the speaker’s in trouble and the next part of the message might not be what you predicted.’”84

B. Filler Utility

Significantly for lawyers conveying complex information, these filler functions, either individually or in combination, demonstrably increase listeners’ memory and comprehension. The mechanisms by which this happens are complex. But simply put, verbal fillers make listeners pay attention, though often (and ideally) on an unconscious level. Because of our shared knowledge about communication, a listener who hears a speaker use a verbal filler knows that the speaker has encountered a disruption in the speech-planning process. That information causes the listener to be more attentive. That heightened attentiveness, in turn, can help the listener to better predict, understand, and remember the information that follows uh or um.

This view of verbal fillers has been demonstrated in a variety of ways. Martin Corley, Lucy J. MacGregor, and David I. Donaldson, at the Universities of Edinburgh and Stirling, measured the neural activity of listeners and found a “profound” difference in the speed at which they processed words that were preceded by a verbal filler.85 This effect persisted over time. Even after a delay of up to fifty-five minutes, listeners were better able to remember words preceded by a verbal filler.86 Jean E. Fox Tree, a professor of psychology at the University of California, Santa Cruz, obtained similar results with uh, which helped listeners recognize words in upcoming speech.87 Fox Tree asked listeners to press a buzzer when they heard a specific “target” word, which they had committed to memory. Listeners responded faster when the target word was preceded by uh than when the uh had been excised.88 A team of linguists at the University of Rochester demonstrated that disfluency overall is a cue to listeners that the speaker is referring to new information.89When used as a correction, verbal fillers have been shown to help listeners process and understand word repairs more quickly.90

Studies show that these benefits in comprehension apply not only to words, but to entire narratives—an important consideration for lawyers conveying their clients’ stories. In a study conducted by Scott H. Fraundorf and Duane G. Watson at the University of Illinois, groups of listeners heard stories paraphrased from Lewis Carroll’s Alice in Wonderland. 91 Some of the stories were told fluently—that is, without verbal fillers or pauses—and other stories were marked by uhs and ums. After hearing different versions of the stories—some told fluently, some spliced with verbal fillers, and some interrupted by coughing—listeners were asked to verbally recall the stories in as much detail as possible.92 Listeners were consistently better able to remember the stories with the fillers.93 This effect occurred regardless of whether fillers were inserted at key plot points, where they might naturally appear, or randomly sprinkled throughout the story.94

Some researchers suggest that, in addition to increasing comprehension by sharpening the listener’s attention, the placement of verbal fillers before new thoughts or ideas also helps to organize spoken language for listeners and give it a type of structure.95 Some linguists go even further and argue that verbal fillers are necessary elements in spoken discourse and that removing them for the sake of an ideal of fluency actually shortchanges listeners.96 In support of this theory, linguists point to the comprehension problems listeners encounter when written work is read aloud. As the late linguist Göran Kjellmer explained, “A lecture that is read aloud from the written page is often difficult to take in when its delivery lacks the verbal guides and signposts that we more or less subconsciously expect to find in speech.”97 Consequently, listeners who hear text read aloud “are in danger of missing the point of the argument.”98 Citing the essayist Louis Menand, O’Connell and Kowal go so far as to describe writing as an inferior form of communication, precisely because it lacks the disfluencies that mark our speech.99 To them, writing is a hieroglyph, while speaking is “a symphony.”100

If the forensics of verbal fillers were the whole story, it would end here. We know that most, if not all, speakers use verbal fillers. There is strong evidence that listeners are better able to understand and remember messages that include verbal fillers. Yet, from the time we are able to speak, it seems, we are exhorted to eliminate uhs and ums from our speech. Why are verbal fillers so despised, and what should speakers do about them?

____________________

1 E-mail from Elizabeth Shriberg to Barbara Gotthelf (Sept. 27, 2006) (copy on file with the author).

2 Id.

3 Id.

4 Jean E. Fox Tree, Discourse Markers across Speakers and Settings, 4 Lang. & Linguistics Compass 269, 272 (May 2010); Susan E. Brennan and Maurice Williams, The Feeling of Another’s Knowing: Prosody and Filled Pauses as Cues to Listeners about the Metacognitive States of Speakers, 34 J. Memory & Lang. 383, 391 (1995).

5 Fox Tree, supra n. 4, at 276–77.

6 Scott H. Fraundorf & Duane G. Watson, The Disfluent Discourse: Effects of Filled Pauses on Recall, 65 J. Memory and Lang. 161, 161 (2011).

7 Nicholas Christenfeld & Beth Creager, Anxiety, Alcohol, Aphasia, and Ums, 70 J. Personality & Soc. Psychol. 451, 454 (1996).

8 Heather Bortfeld, Silvia D. Leon, Jonathan E. Bloom, Michael F. Schober & Susan E. Brennan, Disfluency Rates in Conversation: Effects of Age, Relationship, Topic, Role, and Gender, 44 Lang. & Speech, 123, 125 (2001) (citing Elizabeth Shriberg, Disfluenices in Switchboard, Proc. Intl. Conf. on Spoken Lang. Processing, Addendum, 11 (1996)).

9 Bortfeld et al., supra n. 8 at 128, 141–42. Overall, men used 3.04 verbal fillers per 100 words compared to a rate of 2.07 for women. Men also had more word repeats (e.g., just on the left left side), at 1.74 to 1.21 per 100 words. Id. at 141.

10 Herbert H. Clark & Jean E. Fox Tree, Using Uh and Um in Spontaneous Speaking, 84 Cognition 73, 74 (2002).

11 Fraundorf & Watson, supra n. 6, at 161.

12 Clark & Fox Tree, supra n. 10, at 74.

13 Id.

14 Id. at 78 (emphasis in original).

15 Daniel C. O’Connell & Sabine Kowal, Communicating with One Another: Toward a Psychology of Spontaneous Spoken Discourse 128 (2008). The schwa sound is the a sound initializing and ending America. Merriam Webster’s Collegiate Dictionary 1111 (11th ed. 2005).

16 Clark & Fox Tree, supra n. 10, at 92.

17 Michael Erard, Um . . . : Slips, Stumbles, and Verbal Blunders, and What They Mean 142 (2008). The sign for um is an open palm, with five fingers slightly apart, and a repeated circling of the forearm away from and toward the speaker. Id.

18 Clark & Fox Tree, supra n. 10, at 97.

19 Id.

20 Göran Kjellmer, Hesitation. In Defence of Er and Erm, 84 English Stud. 170, 172 (2003).

21 See Gunnel Tottie, Uh and Um as Sociolinguistic Markers in British English, 16 Intl. J. of Corpus Linguistics 173, 192 (2011). Michael Erard cautions that such conclusions should not be viewed as a measure of intelligence, but instead reflect the norms of one’s community. Erard, supra n. 17, at 100.

22 See Tottie, supra n. 21, at 192; see also Bortfeld, et al., supra n. 8, at 139.

23 Tottie, supra n. 21, at 192.

24 See e.g. Nicholas Christenfeld, Does it Hurt to Say Um?,19 J. Nonverbal Behavior 171, 178–80 (1995).

25 Stanley Schachter, Nicholas Christenfeld, Bernard Ravina & Frances Bilous, Speech Disfluency and the Structure of Knowledge, 60 J. Personality & Soc. Psychol. 362, 362 (1991); see also Clark & Fox Tree, supra n. 10, at 87–88.

26 Christenfeld, supra n. 24, at 172.

27 Schachter et al., supra n. 25.

28 Id. at 364.

29 Id. at 365.

30 Id. at 362.

31 See id.

32 Id. at 365.

33 Nicholas Christenfeld, Options and Ums, 13 J. Lang. & Soc. Psychol. 192, 197 (1994).

34 Id.

35 Id. at 197–98.

36 Id. at 198.

37 Id.

38 Christenfeld & Creager, supra n. 7, at 451–52.

39 Id.

40 Id. at 451.

41 Id. at 452.

42 Id.

43 Id.

44 Id. at 453. The “creative” speakers used an average of 7.03 ums per minute, compared to 4.07 ums per minute in the grammatical coding group.

45 Id. at 454–55.

46 Id. at 456.

47 Id.

48 Id. at 457–58.

49 Id. at 458–59. The study authors warn, “Before suggesting intoxication as a strategy to concerned public speakers, it should be noted that, to eliminate the average speaker’s ums, about 19 drinks in the course of an evening are required.” Id. at 457.

50 Id. at 458

51 Id. at 459.

52 Id. at 452; 458 –59.

53 Id. at 459.

54 Id.

55 Id.

56 Id.

57 Clark & Fox Tree, supra n. 10, at 74.

58 Bortfeld et al., supra n. 8, at 124.

59 O’Connell & Kowal, supra n. 15, at 126.

60 Id.

61 Tottie, supra n. 21, at 193.

62 O’Connell & Kowal, supra n. 15, at 128.

63 See Erard, supra n. 17, at 134.

64 Susan Brennan & Michael F. Schober, How Listeners Compensate for Disfluencies in Spontaneous Speech, 44 J. Memory & Lang. 274, 275 (2001).

65 See O’Connell & Kowal, supra n. 15, at 130–31; see also Karl G.D. Bailey and Fernanda Ferreira, Disfluencies Affect the Parsing of Garden-Path Sentences, 49 J. Memory & Lang. 183, 184 (2003).

66 Christenfeld, supra n. 24, at 178.

67 Id. at 180.

68 See Bailey & Ferreira, supra n. 66, at 184.

69 Id. at 184–85.

70 Kjellmer, supra n. 20, at 182–90.

71 Id. at 183.

72 Clark & Fox Tree, supra n. 10, at 90.

73 See discussion supra at sec. II (A, B). Surveying the literature, Clark and Fox Tree identify the following bases for delay: (1) The speaker is experiencing a planning problem; (2) the speaker is searching memory for a word; (3) the speaker is hesitating about something; (4) the speaker is in doubt or uncertain about something; (5) the speaker is engaged in “speech-productive labor,” such as deciding what to say or how to say it. Clark & Fox Tree, supra n. 10, at 90.

74 See Kjellmer, supra n. 20, at 183.

75 Id. at 183–86; see also Clark, Fox Tree, supra n. 10, at 89–90.

76 Kjellmer, supra n. 20, at 185. Kjeller notes that in some instances, “One can almost hear the voice of the speaker trailing off at the end, hoping to be relieved,” as in the sentence, “I don’t know, I mean er[,] er . . . . ” Id. at 185–86.

77 Clark & Fox Tree, supra n. 10, at 89–90. Clark and Fox Tree cite to the following exchange, in which the speaker (Sam) repeatedly invites the listener (William) to interject. William does so only after the third uh:

 Sam: [B]ut the whole object of this, is to talk about, . . . first, naturally the department, . . . but but also if anybody wants to raise anything else about the college, . . . uh, do please do so, . . . I mean it’s abs- total free for all, . . . um . . . how about things generally, I mean have you uh let’s start with the accommodation, . . . obviously this is a problem, . . . uh . . .

William: I think it’s a problem . . . . Id. at 90 (ellipses substituted for marks designating internal pauses of different lengths). Here, “Sam uses uh and um not to hold the floor, but to signal his willingness to give it up.” Id.

78 Kjellmer, supra n. 20, at 186.

79 Id. at 187.

80 Id.

81 Id.

82 Id. at 188.

83 Id. at 189.

84 Martin Corley & Oliver W. Stewart, Hesitation Disfluencies in Spontaneous Speech: The Meaning of um, 2 Lang. & Linguistics Compass 589, 602 (2008).

85 Martin Corley, Lucy J. McGregor & David I. Donaldson, It’s the Way That You, Er, Say It: Hesitations in Speech Affect Language Comprehension, 105 Cognition 658, 667 (2007).

86 Id.

87 Jean E. Fox Tree, Listeners’ Uses of Uh and Um in Speech Comprehension, 29 Memory & Cognition 320, 324 (2001).

88 Id. Interestingly, Fox Tree found an increase in recognition of words preceded by uh, but not um.

89 Jennifer E. Arnold, Michael K. Tanenhaus, Rebecca J. Altmann & Maria Fagnano, The Old and Thee, uh, New: Disfluency and Reference Resolution, 15 Psychol. Sci. 578, 581 (2004).

90 Brennan & Schober, supra n. 64, at 293. Brennan & Schober caution that listeners are better able to understand fluent speech. Where, however, a word interruption or repair occurred, a correction that included a filler enhanced comprehension. Id. at 293, 295.

91 Fraundorf & Watson, supra n. 4.

92 Id. at 165.

93 Id. at 166.

94 Id. at 170.

95 See Kjellmer, supra n. 20, at 190.

96 Id.

97 Id.

98 Id.

99 O’Connell & Kowal, supra n. 15, at 222.

100 Id.