The Frequency Illusion

Those of us who have bought a car will notice a curious phenomenon where, just as soon as we purchase our car, we immediately begin to see the same model everywhere on the streets. It's remarkable how we all become trendmakers as soon as we buy a car; we singlehandedly turned a Toyota Corolla into a bestseller. After all, we only start seeing this car in mass numbers after we purchase it.

Or perhaps you relate more to having psychic powers capable of influencing the world. You learn about a particular term, or meet someone from a country, and suddenly you see reference to that country everywhere. Once I started studying the Scottish accent more, I found references to Scotland everywhere - Scottish bands on my Spotify auto-generated playlists, Scottish actors referenced on Twitter, and even more books on Scotland at my local bookstore. Could it be that me learning about Scotland raised the prominence of this northern country?

Luckily (or not), these were not incidences of the Toyota Corolla and the country of Scotland suddenly becoming more popular after one learns about them. These are examples of what is popularly called the Baader-Meinhof effect, also called the Frequency Illusion. This is a cognitive bias resulting from selective attention (you don't pay the same amount of attention to everything all the time) and confirmation bias (looking for/remembering things that support our hypothesis and ignoring contrary evidence). The Frequency Illusion is especially relevant in linguistics research, and it was indeed coined by a linguist. We're going to take a look at the Frequency Illusion and some examples of how it applies in language.

History

Arnold Zwicky, a linguist at the University of Stanford and former writer for Language Log, codified this observation in a 2006 post.

In any case, we have here another instance of the Recency Illusion, the belief that things YOU have noticed only recently are in fact recent. This is a selective attention effect. Your impressions are simply not to be trusted; you have to check the facts. Again and again -- retro not, double is, speaker-oriented hopefully, split infinitives, etc. -- the phenomena turn out to have been around, with some frequency, for very much longer than you think. It's not just Kids These Days.

Professional linguists can be as subject to the Recency Illusion as anyone else. Charles Hockett wrote in 1958 (A Course in Modern Linguistics, p. 428) about "the recent colloquial pattern I'm going home and eat", what Laura Staum has been investigating under the name (due to me) the GoToGo construction. Here's an example I overheard in a Palo Alto restaurant 8/6/05: "I'm goin' out there and sleep in the tent." But Hockett's belief that the construction was recent in 1958 is just wrong; David Denison, at Manchester, has collected examples from roughly 30 years before that.

Another selective attention effect, which tends to accompany the Recency Illusion, is the Frequency Illusion: once you've noticed a phenomenon, you think it happens a whole lot, even "all the time". Your estimates of frequency are likely to be skewed by your noticing nearly every occurrence that comes past you. People who are reflective about language -- professional linguists, people who set themselves up as authorities on language, and ordinary people who are simply interested in language -- are especially prone to the Frequency Illusion.

Here at Stanford we have a group working on innovative uses of all, especially the quotative use, as in the song title "I'm like 'yeah' and she's all 'no'". The members of the group believed that quotative all was very common these days in the speech of the young, especially young women in California, and the undergraduates working on the project reported that they had friends who used it "all the time". But in fact, when the undergrads engage these friends in (lengthy) conversation, tape the conversations, transcribe them, and then extract occurrences of quotatives, the frequency of quotative all is very low (quotative like is really really big). There are several interpretations for this annoying finding, but we're inclined to think that part of it is the Frequency Illusion on our part.

Zwicky wasn't the first to notice this phenomenon; the name Baader-Meinhof Effect had been circulating on message boards since 1994, and is more widely recognized. This wasn't even the first time Zwicky had written about it; he noticed his own sensitivity to a character in a book saying "and yet" despite the phrase only occurring three times in a novel. He has made a collection of posts on the Frequency Illusion, helpfully listed on his site.

He uploaded a short pamphlet to the Stanford University website that helpfully summarizes the sub-types of the Frequency Illusion:

Why are many people – including, on occasion, linguists – inclined to systematic dogged misapprehensions about variation in language (like the five below, from postings to the Language Log over the past two years)? These illusions follow from psychological processes and social practices, combined with bits of language ideology, and are facilitated by the fact that hardly anyone has a panoptic view of language variation; we mostly have to think about it on the basis of our personal experience.

(1.1) Recency Illusion: If you've noticed something only recently, you believe that it originated recently. (Example: a widespread belief that the case selection in between you and I is a recent innovation.)

(1.2) Antiquity Illusion: If you do something yourself, you believe that you always have, and that people in general have done it for a long time. (Example: the idea that the whole nine yards is a venerable idiom.)

(2.1) Out-Group Illusion: Things you view as novel, or simply ba d, are characteristic of groups you don’t see yourself as belonging to. (Example: an Australian’s assertion that “double is” is not found in Australia, but is probably an Irish thing.)

(2.2) In-Group Illusion: Things you view as characteristic of groups you see yourself as belonging to are peculiar to those groups, not shared by outsiders. (Example: many Pittsburghers’ belief that items like needs washed and redd up are unique to their city.)

(3) Frequency Illusion: Once you notice a phenomenon, you believe it happens a whole lot. (Example: a widespread belief that quotative all occurs “all the time” in the speech of some (young) people.)

One of the most humbling observations that Zwicky has made is that the Frequency Illusion can happen to anyone, even to professional linguists. The reality is that there are far too many things happening in language for one to be able to keep track of. Even keeping up with published papers is a challenge, let alone running statistical models of every phrase that catches the ear. It is easy for us to be mislead as to how common something is, and our intuitive responses on who is doing it are often simply wrong.

That being said, these intuitive responses can be a good jumping off point. I often go into my Dialect Dissections with certain expectations about what I will find and which features will be the most salient, but almost every time the evidence takes me in a different direction from what I expected. That being said, without my initial intuition, I would not have even known where to look. When the Frequency Illusion is the beginning and the end, it can be harmful. But if we let it guide research, we can learn what is really going on - and then hopefully share with others to 'shatter' their own Frequency Illusion.

Although Zwicky has already discussed several examples of the Frequency Illusion in his writings, I'm going to extend that analysis to some other areas.

Examples

One of the most noticeable examples of the Frequency Illusion is the discourse surrounding creaky voice, also known as glottal fry and vocal fry. Creaky voice is a kind of phonation that results in a noticeably 'croaky' or 'creaky' sound. It is an important phonemic mode in some languages, where presence and absence of creaky voice tells some words apart. But most of the attention paid to creaky voice isn't about its use in distinguishing words. Instead, it has focused on the supposed epidemic of creaky voice affecting college age girls in the United States.

Carrie Gillion has a great article discussing creaky voice. She mentions that creaky voice used to be associated with British men, up until an article was released that mentioned that college age women use creaky voice often. At this point, the association flipped and also took on an increasingly negative tone.

Gillon mentions that prejudice against young female speech is an important part of this. I would like to add that this is an example of the Frequency Illusion, for I have found a surprising number of examples of creaky voice from before 2011 that nobody seems to have been bothered by.

An example of the old-fashioned association of creaky voice with British men can be found in "Courage the Cowardly Dog." The character "Naughty Fred" appears in a 2005 episode. He has an English accent, although he is played by a Canadian actor. What is curious is the amount of vocal fry that his actor applies. It is especially noticeable each time he says 'naughty.' This episode aired in 2005, before the hullabaloo about creaky voice. Perhaps his voice actor still associated creaky voice with English men. I haven't found any references or comments about the high amounts of creaky voice this character displays. The notion that creaky voice is a new phenomenon or only prevalent in young women thus seems to suffer from a mass example of the Frequency Illusion striking at a particular time.

"Cause me to be quite ~naughty~"

Creaky voice as an artistic choice has also been criticized heavily. On two different ends of the pop music spectrum, we have Britney Spears, who was hugely influential in popularizing the use of creaky voice, and on the other end we have indie and indie-aligned artists like Billie Eilish who also use creaky voice frequently. Both usages of creaky voice are subject to heavy criticism and attack - though perhaps not that heavy since both Britney Spears and Billie Eilish are successful singers who have sold millions of records.

Creaky voice as an artistic effect is not even new, although its increased prevalence may be. In the song "Can't Smile Without You," we hear Karen Carpenter use creaky voice to transition from the musical phrase "finding it hard even to talk" to "and I feel sad." Karen Carpenter is the poster child for that pure, 'clean' 70s style of singing, with few affectations and a nice dose of vibrato on sustained notes. It is surprising to hear her use creaky voice in much the same way Britney does, to demarcate different musical phrases. While her use is definitely very limited compared to 1990s+ artists', it damages the notion that creaky voice is a novelty.

"~I~ don't even talk to people I meet ~and I~ feel sad when you're sad"

We can continue finding examples of the Frequency Illusion. For instance, I highly suspect that some aspects of Indie Voice are subject to the Frequency Illusion. The musical style is very common, but the linguistic style is rarer. The diphthongization, which was the subject of a Buzzfeed article, is difficult to find consistently. My articles on Indie Voice collect a number of these examples. There are even quotations from people complaining about how 'these artists are saying dreyss' and how you 'have to add an i' to sing in this style. But many of the early examples are noticeably subtle. Indeed, I have had several commenters tell me that they cannot hear the supposed diphthongization in the early examples.

There is evidence that something similar to this diphthongization does happen in spoken speech, as reported by John Wells and in the scant two examples I found. But if this has been happening since 1980, the publication date of the John Wells book, why did people only start talking about it now? There are pre-2013 examples of this subtle diphthongization, but there are few online records of people specifically talking about the diphthongization. I even have had commenters tell me that they do something similar in their own speech and that they don't consider it a noticeable linguistic phenomenon. If this has been happening in speech and in song, why did people start paying attention?

The indie voice example does not have a clean resolution. Perhaps the Frequency Illusion caused people to associate otherwise uninteresting phonetic features specifically with the singing style. I don't think the Frequency Illusion is the whole story, but I would not be surprised if it weren't a part of this perception that 'indie voice is everywhere' when actual linguistic examples are not as easy to come by.

Conclusion

I once read a post on a message board where a man in his 50s was complaining about how 'young people' cannot speak English correctly. His example was the word 'tree', which he claimed young people pronounced 'chree.' It seems this man was unaware that this sort of affricative assimilation is common not just in English, but cross-linguistically.

He had likely heard this form many times in his life and never paid it any mind. But when it came from the mouth of someone he was suspicious of - a young person - his brain suddenly noticed it and not only registered it as a 'young person thing,' but started hearing it everywhere and taking it as proof that 'this is how all young people talk.' It was his subjective experience, but his conclusion was not based on any evidence.

We all need to be aware that our intuitions often lead us astray when it comes to linguistic phenomena, and that pop explanations can lead us astray. Our own feelings about a group (young people, young women, the Irish, Pittsburghers) can cause us to inaccurately attribute certain features as being unique to them and as also being newer than they seem. If we want to have an understanding of language variation that lines up with reality, then we need to as much as possible collect evidence to try and test these intuitions. We also need to be wary and skeptical of 'new' features without accompanying proof, especially if we are claiming some particular group is doing it.