This Post Is About, Umm... Disfluency

I'm a slow talker. If you were being kind, you might describe the pace of my pace as "pensive." If you were being accurate, you might describe it as "bumbling." As a result of my slowness, I say "uh," and especially "um" a lot. A whole lot. So much, in fact, that my son frequently demands out of frustration that I "stop saying um!" In the literature, such verbal fillers, and the related, "er," "I mean," "you know," and in children from the 80s and the Aughts, "like," are mercifully referred to as disfluencies, or even more mercifully as "performance additions." And over the last decade, psycholinguists (not a very merciful name) have found that they may actually play important roles in speech.

Traditionally, "uh" and "um" were thought to be involuntary products of a momentary difficulty in processing what one wants to say, or in deciding whether one actually wants to say it, and therefore are meaningless themselves. Alternatively, they were seen as merely a means of preventing people from interrupting during a pause in speaking. There's an obvious problem with this view, though. Why do we have more than one such marker of disfluency? In fact, English isn't the only language with more than one. Clark and Fox Tree¹ found equivalents of both "uh" and "um" or similar fillers in multiple languages, including Japanese, Spanish, Norwiegan, Swedish, Dutch, French, German, and Hebrew. Distinct elements with no differences rarely survive in a language, much less several languages from different families, so there must be something to "uh" and "um."

It's Gotta Be the Shoes!

I remeber that when I, however rarely, got the "hot hand" in a game, my basketball coach would yell "it's gotta be the shoes!" He knew I had the hot hand, and he'd tell my teammates to get me the ball, but he couldn't attribute the hotness to me, so he joked it must be the shoes.

The truth, of course, is that I didn't have the hot hand, and neither did my shoes. That's because there's no such thing as the "hot hand." In a classic paper, Gillovich et al.¹ showed this by looking at every shot taken by the 1980-81 Philadelphia 76ers in their 48 home games (including the playoffs). They found that making a shot didn't increase the probability of making the next shot. In fact, it decreased that probability slightly. In other words, Dr J was slightly less likely to make a shot if he'd made his last shot! Furthermore, looking at the "runs" (streaks of misses or makes) showed no evidence of a grouping of makes or misses above what you would expect by chance.

Why does pretty much everyone believe in the hot hand if, in fact, it doesn't exist? Well, because our brains are not very good at figuring out probabilities or perceiving randomness, and they're primed to see patterns where they don't exist, so all it takes is a small run to make us think something good is going on. Or as Gillovich et al. put it (p. 311-312):

Evidently, people tend to perceive chance shooting as streak shooting, and they expect sequences exemplifying chance shooting to contain many more alternations than would actually be produced by a random (chance) process. Thus, people "see" a positive serial correlation in independent sequences, and they fail to detect a negative serial correlation in alternating sequences. Hence, people not only perceive random sequences as positively correlated, they also perceive negatively correlated sequences as random... We attribute this phenomenon to a general misconception of the laws of chance associated with the belief that sall as well as large sequences are representative of their generating process. This belief induces the expectation that random sequences should be far more balanced than they are, and the erroneous perception of a positive correlation between successive shots.

More recently, Yigal Attali, in a paper in press, shows that even people who spend their entire lives around basketball can't get past the "hot hand" belief. Here is the abstract:

Although “hot hands” in basketball are illusory, the belief in them is so robust that it not only has sparked many debates but may also affect the behavior of players and coaches. On the basis of an entire National Basketball Association season’s worth of data, the research reported here shows that even a single successful shot suffices to increase a player’s likelihood of taking the next team shot, increase the average distance from which this next shot is taken, decrease the probability that this next shot is successful, and decrease the probability that the coach will replace the player.

So the belief in the hot hand causes players to not only be more likely to take the next shot, but to take more difficult shots just because they made their last shot, even though they're more likely to miss that second shot. Despite this, coaches are so convinced of the hot hand effect that they're less likely to punish players for the resulting poor decisions.  Again, what's amazing about this is that these people are experts in basketball, and have witnessed many thousands of in-game combinations of shots, but they're still unable to perceive that making one shot doesn't make it more likely that the same player will make the next shot, especially if that next shot is an even more difficult one.

I blame the shoes.

1 Gilovich, T., Vallone, R., and Tversky, A. (1985). The hot hand in basketball: On the misperception of random sequences. Cognitive Psychology, 17, 295-314.

About the Blog's Title

This blog is going to be (mostly) about cognitive psychology, so I wanted a title that was a nod to something, or someone, important to the field. After a bit of deliberation, I decided Hermann Ebbinghaus was a pretty good candidate, but Ebbinghaus is kind of a weird blog title, so I went with his most famous finding, the forgetting curve. You can easily find short descriptions of Ebbinghaus’ life and work online, but I thought I’d say a little bit about why he’s important. By 1885, when Ebbinghaus published his classic work usually translated into English as Memory: A Contribution to Experimental Psychology (which you can read in its entirety here), the study of the human mind or soul was already a couple millennia old at least, but to that point had primarily been studied via reasoning from introspection (even Wundt, generally considered the father of modern psychology, used an introspective method). For some time, psychophysical phenomena such as vision and audition, had been studied using the methods of natural science, because these phenomena were thought to fall under the purview of biology. Higher-order mental phenomena like memory or reasoning, on the other hand, were, if not of a different metaphysical sort than the biological, at least not amenable to the same methods of investigation. Ebbinghaus (and others) thought otherwise, and he set about to both demonstrate that the scientific method could be applied to cognition, and to learn something about memory in the process. So he says in the preface:

In the realm of mental phenomena, experiment and measurement have hitherto been chiefly limited in application to sense perception and to the time relations of mental processes. By means of the following investigations we have tried to go a step farther into the workings of the mind and to submit to an experimental and quantitative treatment the manifestations of memory.

So using only one experimental subject, himself, Ebbinghaus set about memorizing lists of nonsense syllables and, to determine how memory for familiar and meaningful syllables might differ, six stanzas of Byron’s Don Juan. He explored memory from several different perspectives, looking at how quickly syllables of different length are learned, how his memory for the syllables was affected by how long he studied them (SPOILER: studying them longer makes you remember more for a longer period of time), how repeatedly learning the lists affected memory, and how the order of the syllables influenced memory. The forgetting curve, however, comes from his exploration of the effect of time on retention.

Here’s what he did, in short: he first learned the syllables, or stanzas, until he could repeat them all in order perfectly. Then he’d wait for some period of time and relearn them. He measured his retention by comparing the number of times it took him to relearn the list perfectly to the number of times it had originally taken him to learn it. The fewer times it took him, the more work his retention of the list had saved him, so he called the measure “savings.” If he was able to recall the list perfectly on his first try after the delay, retention was 100%. If it had taken him 10 times to learn the list the first time, and he recalled it in 2 tries the second time, then retention was 80%. If it took him 8 tries, retention was 20%. And so on (the complete numbers are in the table at the top, taken from his book).

The results were pretty simple: between the initial learning and a test of savings a short time later, there was large drop in retention, from the 100% after the list had initially been learned perfectly to 58.2% after just 20 minutes. After that, the reduction in savings became smaller with each interval, so that the difference between savings after 3 days (25.4%) and 6 days (21.1%) was much smaller than the difference between the initial learning and 20 minutes. This pattern gets us this graph (via, who apparently got it from Purdue University):

And that is what the basic Ebbinghaus forgetting curve looks like. It seems simple now, and pretty obvious too, but remember this was produced in a time when quantitatively measuring memory was thought by many to be impossible. This was, however seemingly mundane to us, revolutionary to the psychologist of 1885. Cognitive psychology has come a long way since then (for one, we don’t use ourselves as subjects, and we almost always use more than one), but these are the field’s not so humble beginnings. And so in lieu of naming the blog Ebbinghaus, honoring the discipline’s father with the name of his most famous finding seems appropriate to me.