12
$\begingroup$

[Reraising a recently asked question that was closed here, but in more cogsci fashion, I hope.]

Some people are more inclined than others to omit a word when typing, but still be unaware of its unintended absence on a (first) proofreading. In human-computer interaction (HCI) this is called a "misperception error". (according to Jan Noyes, Designing for Humans, p. 147, Psychology Press 2002... but that's about all the book says on the topic).

So what else is known?

  • First of all, is it also called something else, perhaps more specific, because "misperception error" is very general, so it's hard to turn up any research on this particular error using these terms in pubmed.
  • What is the prevalence among general population? It can't be something that hard to study, say among students, the low-hanging fruit of psychology and HCI studies, so I suspect some study on prevalence exists.
  • What are some abnormal psychology correlates? I would venture a guess dyslexia and ADHD, but how much more common is it in such sub-populations than in neurotypicals?

I would gladly break this into separate questions if deemed necessary, but I suspect all of the above can probably be answered from one or two papers specifically on this topic.


I found a paper on a similar auditory phenomenon, which has some generally applicable ideas:

the application of signal detection theory (SDT) allows perceptual performance to be understood in terms of both sensitivity (the ability of a system to distinguish signal from noise) and response bias (the system’s bias, independent of sensitivity, for responding either positively or negatively)

So obviously there is a stronger response bias in proofreading your own typing than another's... which probably means that general statistics for "filling in" missing words in a sentence need to upward adjusted for proofreading your own typing.

Even more interesting, and perhaps a little worrisome for those exhibiting this,

This theory is supported by behavioural evidence showing that the influence expectation exerts over auditory perception is stronger in those who hallucinate [...] This theory is also supported by evidence that the prediction error signal, which communicates the discrepancy between expectation and actual sensation [...] is reduced in those who suffer from schizophrenia [...].

It remains to be seen if misperception of non-existent words due to semantic expectation correlates to some kind of predisposition for visual or other kind of hallucination. It seems a bit more far fetched.


Perhaps it's also worth noting here that simply proofreading familiar text, which the subject has read before in the correct version, but which is then modified by the injection of spelling errors, actually improves the detection rate of spelling errors. So not all kinds of familiarity with the text predispose to proofreading errors, at least not to proofreading spelling errors (missing or extra words were not tested in that study).

$\endgroup$
  • 4
  • $\begingroup$ I also found a blog with this bit of info "In discussing notable case histories of aphasia, Sacks mentions the English writer Samuel Johnson, who experienced aphasia after a stroke at the age of 73. While Johnson eventually regained the ability to speak, he “made uncharacteristic mistakes, sometimes omitting a word or writing the wrong word” in his writing and correspondence." I don't think aphasia is the right term for this though... a lot of things can happen in/after a stroke, although the correlation is noted. $\endgroup$ – Fizz Nov 10 '17 at 18:57
  • $\begingroup$ @ArnonWeinberg: thanks for those. The paper found by Alice was the closest to what I'm asking here, except that it's almost entirely focused on word skipping (e.g. missing a duplication like "the the"). I suspect that the word length influences the probability to omit it as well. Omitting is called "“fill in” the gaps" in that paper. $\endgroup$ – Fizz Nov 10 '17 at 19:22
  • 1
    $\begingroup$ @Fizz Are you certain some of your updates would not constitute a partial answer? Do not forget you can (and are encouraged) to answer your own question! :) You can even add multiple if they for example answer from different perspectives. $\endgroup$ – Steven Jeuris Nov 13 '17 at 10:22
2
$\begingroup$

In his 2003 psycholinguistics book, John Field has summarized (pp. 70-72) his own typing errors and combined them with an older corpus of Hotopf.

Missing words were among the frequent errors, but alas no numerical frequencies are given.

But what he says he noticed is that short function words like "are" or "it" (these are his examples) are more likely to be omitted. And regarding the type of words more likely to be omitted he says:

An important factor appears to be the difference between lexical and function words. The latter seem very susceptible to error: they are sometimes omitted or duplicated, and it is quite common in rapid typing for one function word to be substituted for another, even when they are not phonologically identical. This leads one to the conclusion that function words may be processed differently from lexical ones. It seems possible that the writing process accords a higher level of attention to lexical items.

Although he doesn't say this, the full-context examples that he gives for omission seem to have another property, namely that short words get omitted close to each other (compare with duplication)

  • though these [are] not frequent

  • that [it] is possible

And if you're curious about the related issue(s) he mentions, among the frequent duplication examples are "the the", "that that", and "of of". Among the substitutions are "the" instead of "that", "of" <- "for", "doe" <- "do", "were" <- "where", "for" <- "from". He continues:

Finally, notice that some writing slips provide evidence of a planning process operating at a level higher than the word. When writing a particular word, we already have the next few words stored in our minds, ready for production

The examples for these are more interesting

  • importan topic
  • these given (= give) insights
  • using bothing (= both) approaches
  • difference (= different) intelligence tests
  • neighbourshoods

This is more qualitative than quantitative, but for lack of anything better, I put it in the answer box.

So if Field is correct, there's an anticipatory process in writing that causes a number of spelling errors, not all of them omissions.

$\endgroup$

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.