Mental Lexicon

Update: In response to a matter than came into our family - the Iditarod musher from NH who was going to be staying with my family in Nome had her dog team hit by a pickup in Willow - I dictated "This is not our usual sort of tragedy" when I meant to say charity, and changed it before sending.  Similar sound, with syllables and stress pattern the same, but not similar in appearance. It is interesting that the meanings do appear to be connected, as if the one word were hovering somewhere near my target word. So malapropism by sound, meaning, or both?

******

How we choose the words we speak is fascinating.  There are imbalances in the storage which creates mistakes. You know how it is when you are searching for a word. You rummage around in your brain looking for clues.  If you are with someone who might know the answer you say things out loud. "It's like decrepit but it's not so old-fashioned.  It's a newer word." That would be searching according to meaning. Or "It begins with a 't' and it's three or four syllables. I think it's a foreign word." That is searching according to sound and impression. In the category of sound are also those things like whether it is foreign, what feel it has, and we nearly always know exactly what part of speech it is as well.

We have a bias toward choosing by sound, which is how malapropisms and related mistakes happen. The malapropisms found in literature are not quite the ones we create in real life. Those are created slyly, with an eye to humor, as the original Mrs. Malaprop in Sheridan's 18th play "The Rivals" was, saying "allegory" instead of "alligator." In literature, or these days in the gotcha moments of social media, emphasis is placed on misuse of a similar sounding word, with the usual implication that "this stupid uneducated person is trying to hard to show off his/her vocabulary, with hysterical results which show that our opponents are stupid and we, by extension, are the Smart People." It is equal opportunity, anyone can play. The first example given is Archie Bunker saying "We need a few laughs to break up the monogamy." But the belief seems to come from the memorability of humorous mistakes plus our experience of people misusing words in general.  There actually doesn't seem to be evidence that these are people reaching to look smarter than they are. They are just mistakes, and four sentences later the person may use the word correctly. An exception would be people speaking English as a second language. All peoples have some tendency to regard those who speak their language less well as less intelligent and make fun of inaccuracies, but we mostly know this is not really the case, just an impression that should be overridden.

We like studying how things go wrong, because the suggest possibilities of how the whole operation works. There are other errors related to malapropisms, which also tell us something. There are are anticipation or perseveration errors, like saying "splacing from one tape" instead of splicing because your brain sees the important long "a" sound coming*, or "pale skay" instead of sky because you are residually stuck on the long "a." There is also an inadvertent mixing, such as being caught between saying jumping and leaping becoming "lumping," or for humorous effect such as Lewis Carrol's "Jabberwocky," where slither and slimy become "slithy." We see those for what they are, just stumbles from hurrying or half-thinking of something else. Sometimes they are funny, especially when the error is also a word, but we seldom think the person stupid, just inattentive. There are also spoonerisms, clearly a sound processing error not a misunderstanding of meaning. 

Freudian slips are not common. At most they may reveal a word we have hanging around too near the surface that we should stuff down a little better, but symbolic meanings are so often explained by other types of sound errors that we cannot rely on the psychoanalytic explanation as telling us much.  I would love to believe that Ted Kennedy's statement that"the purpose of education is to encourage the breast and the brightest," means he can't think of education without wanting to undress schoolgirls, but it is more likely he just has "breast" too close to the surface.

A study by David Fay and Anne Cutler,  Malapropisms and the Structure of the Mental Lexicon, shows how indirect evidence from speaking errors can reveal something about the structure of human communication. I should note that the study dates from 1977 and I don't know if it has been challenged, overturned, or replicated.

They use the phrase "Speech production device" to mean the part of the brain that organises this.  They are not speculating on what this "device" is in the brain. They say they are not taking sides, and the way linguistics arguments were going at the time, they perhaps worried that readers would get distracted by their apparent support for one model or another.  They are sidestepping that issue saying "Well, whatever it is up there, it has to be doing this in some way."  Good move, I think, especially as they say that transformational grammar has at least one problem as a full explanation.  that itself would have boiled some blood in those days.  Maybe now, too.

The analogy they come up with is that our lexicon is like a dictionary in one sense, organised by sound, which is similar to spelling. Yet we have other methods of organising in order to find the word we want. But even if we are able to find words by meaning when we want to say something - how else would we talk? - it does not follow that we have primarily organised them that way. It seems to be secondary.    "...a pairing of sound and meaning must be used in both production and comprehension of speech." Yet the retrieval of the word has to be reversed for production and comprehension.  We are trying to find the words to say according to meaning.  But we first hear the sound when we are retrieving. So the dictionary must work for both. A duplicate dictionary with every entry accessed in both ways would be evolutionarily inefficient. I suppose we would then not have the linguistic errors noted above, but they aren't that frequent anyway. One would think that assembling the dictionary, the lexicon of our communication, would be about equally dependent on sound and meaning, yet it appears that sound is deeply favored in how we create the lists. That the speech errors above are all sound errors is evidence of that. Errors in meaning are often nuanced or subtle, such as "I think uninformed would be a better word than ignorant there." 

It would seem intuitive that we would favor hearing and speaking equally, so the result is surprising. However, the sounds we hear are more likely to be interfered with, and we often need to sort between sounds, even competing conversations, as opposed to sorting between thoughts in our own head.  When that happens, we can make executive decisions to suppress one train of thought for a moment while we attend to another, even switching rapidly among a few conversations going on at the same time.  But we cannot enforce that on the sounds of the rest of the world. Sound is the more efficient storage.

We are storing by sound, so when we have a meaning we want to convey, we have to go to our lexicon, our dictionary, and pull something out we make a series of deeply-inspired guesses based on our history of word-choice since childhood. But if you are opening the dictionary and moving on you might choose an adjoining word. Malapropisms nearly always start with the same letters.  Interestingly, they are nearly always the same part of speech, have the same number of syllables, and even have the same stress pattern as the target word. So our storage by sound includes those other elements of structure. From the study:

we might expect that near neighbors in the dictionary might be very similar in sound;
we might expect that if the production device, homing in on a particular word in the
lexicon, were to err just a little and pick, instead of its target word, that word's ''next-
door neighbor", or a near neighbor, then the word it chose by mistake might sound
very similar to the target word, but would be unlikely to bear any relation to it in
meaning. We might expect, in other words, to find errors having exactly the
characteristics of malapropisms.

Further evidence would be that retain our comprehension of a language we once spoke (not a language we had to produce some phrases for because of school or travel) well after we can no longer speak it.  "I don't speak French anymore, but I still understand it when my mother gets together with her mother and her aunts." In learning to hear a language one of the primary things we do is divide the sounds correctly into words, even when we don't know all the words. Humans speak quickly and spoken language takes shortcuts all the time, so teasing out the pieces from the run-together sounds is difficult at first.

Even though we would not confuse Counting Crows with Casting Crowns by music, we might confuse them by sound. Notice, however, that they are similar in appearance and structure, which are not quite the same as sound,though related.

My primary mistakes are spoonerisms, which at times in my life I have generated automatically dozens of times a day. I keep them to myself unless they are humorous. Yet I find that I consider them funnier than others do, so I seldom mention them at all now. I thought it was hysterical to tell my sons to eat their keys and parrots.  They disagreed. My granddaughters are not any more impressed with that or other such things. Interestingly, the tendency disappeared decades ago when I was on an SSRI for OCD (I eventually decided the subtle side effects weren't worth it), but I didn't notice the lack.  I only noticed when I discontinued the medication and the spoonerisms returned.  "Huh. I haven't thought of those for years."  My wife makes semantic errors that are rather the reverse of a malapropism:- she will say "left" when she means "right," or "I will make hamburg with meat sauce for dinner," meaning spaghetti with hamburg in the sauce. She does not self-correct after hearing, because she does not hear herself.  If she says a sentence I didn't quite catch and I say "What was the first part of that again?" She will always give me a different set of words with the same or very similar meaning. Or again, if I ask "What is the high temperature supposed to be tomorrow?" She will say "I think it is supposed to rain." It would be a great comic character on a sitcom. We have commented over the years that her word-retrieval is different from other people's, but after enough tip-of-the-tongue plus the errors, we can see the pattern. Most of us have an automatic retrieval of the exact words we have just spoken which fades over time. There is a short-term memory mechanism that may be as much related to hearing our own voice as it does to memory of the words we chose for our meaning. She does not have this. I was quite amazed when I learned this, as I thought everyone had it.

My wife, without that reliance on sound (because part of the device is broken) compensates by storing primarily by meaning.  We previously described that with the analogy of "folders" in a file cabinet, as she often will pick something from the correct category, even if it is wrong, even an opposite. Which likely accounts for her astronomical scores on verbal analogies and inspired puzzle guessing by a letter or two of the word (because the sound does not distract her. She has stored by reading knowledge, not what she has heard). I think I have mentioned that we believe she was the record-holder on the Miller Analogies Test for many years. One more example of how compensating for a "disability" can result in a formidable strength. 

She does have other storage that is not sound but is related because it is the word's structure. She might say "cupboard" instead of "classroom" because they are both two words put together. She will store by number of syllables, but not, I don't think, by stress pattern. But even among the studies that show storage by number of syllables, there are enough exceptions to suggest that some of that may be storing by length - which would also show up in reading a word and storing it that way.

I have mentioned before that there is usually no shame in mispronouncing a word in a way that telegraphs that you have read a word but not heard it in conversation. It could even be something to be mildly proud of.  I seldom mispronounce anything, and our early courtship was marked by Tracy using a word, me gently correct the pronunciation, and then, a few beats later asking "what's it mean?" She still has things she never did learn the right pronunciation of, and I still have words I never have quite gotten the hang of, so it's not foolproof.  But it has been a benefit to both of us. However those mispronunciations can sometimes be revealing, as they demonstrate that you don't move in circles where the word is used, even if you have some idea what it means. There is no shame in a constitutional law professor saying corpse-man for corpsman, though others might roll their eyes as to what that signifies about social circles.  But a commander-in-chief should start hanging out with people who use the word regularly.  Though admittedly, the Joint Chiefs might not use it that much because it belongs to distant ranks and their distant past.

*This is an established principle in linguistics when it comes to pronunciation, that a sound will change because of its adjoining sounds, or even in anticipation of a sound a syllable away. The unvoiced "t" in later is surrounded by voiced sounds and becomes lay-der. But lasted holds its "t" better because the s is adjoining.  (In both cases you will note that this is not entire, and we often pronounce the word a bit more crisply as written when we are in more formal settings.)