When you know something, is that different from remembering? Both types of thoughts are clearly part of the memory system, but is there really any difference between the two concepts? We often use the two terms nearly interchangeably: I might say "I remember Suzanne had her purse when we left the restaurant because I saw her pull out her phone at the bus stop," but I might equally say "I know Suzanne had her purse on the bus because she was gabbing on the phone the whole ride home."
But the subtle linguistic difference between the two terms isn't meaningless. We might know George W. Bush is President, but if we remember that James Garfield was the President who preceded Chester A. Arthur, it's likely due to a specific recollection from 11th grade history class (either that or we attended Garfield High School).
We say that we know something because of a general sense that it is true, but we remember something because we recall a particular incident. Psychologists have actually been able to measure the distinction between the two, simply by asking test subjects whether they know or remember the answer to a question (and explaining what they mean by the two terms).
So how does a "know" memory get formed? Are there circumstances when we're more likely to form a "remember" memory? Can one sort of memory transform into the other? Or is the difference between know and remember simply an artifact of the fact that we remember some things better than others? A team led by John Gardiner had volunteers listen to a set of 60 different words, half of them in a male voice and half in a female voice. They rated each word for clarity and pronounceability on a scale of 1 to 5. Then they played a computerized board game for 20 minutes as a distraction. Finally they were tested on which words they remembered.
The test consisted of 120 words, so listeners had only heard half of them before. They were asked to press a button if they had heard the word before, and also to say whether they knew, remembered, or were just guessing. There was one more trick to the test: Half the words were spoken in the same voice they had been originally, but half were spoken in the other voice (formerly male voices were now female, and vice versa). Here are the results:
As you might expect, when the words were spoken back in a different voice, overall accuracy diminished significantly, but it was entirely due to the "remember" responses -- listeners "knew" just as many words as they had previously. In a final twist, some listeners were given a shorter time to respond -- just 700 milliseconds. How did these listeners do?
While respondents were overall less accurate, the same pattern emerged: significantly fewer "remember" responses and the same number of "know" responses when the words were spoken in a different voice.
So different conditions appear to make it more difficult to recall words we remember, but not words we know.
In a separate experiment, the researchers distracted listeners while they listened to the words. In addition to rating the pronounceability of the words, listeners also watched pairs of numbers on a computer screen, and they had to indicate whether both numbers were odd, both were even, or they were mixed. The test at the end was the same as before. Here are the results:
This time the listeners said they knew significantly fewer words when they were presented in a different voice, while the portion they remembered remained the same. As before, when respondents had to answer quickly, the same pattern held:
Overall accuracy diminished, but again the difference between words in the same voice and a different voice was entirely due to the "know" responses.
In two additional experiments, Gardiner's team replicated the results using pictures instead of words. The researchers argue that these experiments show that the know/remember distinction isn't simply a part of a continuum. This work, they say, suggests that knowing and remembering are qualitatively different phenomena. It's also clear that being distracted at the time a memory is being formed biases the memory towards knowing rather than remembering.
It doesn't necessarily make you more accurate, though. I looked up the "Garfield" answer and found that I must have been very distracted in high school: In my first draft of this post, I had him preceding Teddy Roosevelt.
Gardiner, J.M., Gregg, V.H., Karayianni, I. (2006). Recognition memory and awareness: Occurrence of perceptual effects in remembering or in knowing depends on conscious resources at encoding, but not at retrieval. Memory & Cognition, 34(2), 227-239.
Another thing that could be tested for in this same framework would be to have test subjects go through a list (printed or on computer screen) of words, indicating which of the words were part of their everyday vocabulary, which they were familiar with from reading/school, and which were new or unknown to them. I guess this should probably be done after the testing above, so as not to bias the test results, but I'd guess that words which are part of everyday vocabulary are "known" and less familiar words are "remembered."
I suspect whenever I have read something by a cognitive psychologist about how we know anything, how our brain develops any concept, I've looked at their Venn diagrams and other illustrations and wondered what counterpart such drawings have in reality, meaning anatomy, physiology or even my subjective experience of my own brain.
It occurs to me from this experiment you present here that my preference for reality is that I can remember reality in a way that I can't for something that is purely conceptual. I remember cross-sections through the brain and the functions associated with different structures. I remember patients with some neurological deficiency and the defect on their imaging studies that went along with that. For some aspects of such memories, I have no doubt at all, even though I know that as I cross the line into interpreting those memories, I begin to sense doubt, even though I would claim to know both the memories about which I'm certain and the interpretations that I'm just pretty sure are correct. I might not claim to know them as certainly as I know what I would call most of the colors on this webpage, but it's close.
I don't know if it's experience with science or something else that makes me want to have memories of something real to back up what I know. I've noticed some people value reasoning to support knowledge more than I do and experience or data less than I do. At the same time, some people develop false memories to support what they know. (Did you have one of those about TR after Garfield?)
I find myself searching for memories to support what I might know all the time, such of my memories of reading about cognitive psychology. What am I doing physiologically when I do that? Where and how does our comparison of knowledge and doubt play out? I remember a few times when I had no doubt about something, when my faith was perfect, but I discovered that I was wrong. How did that happen?
Where does my desire for knowledge come from? Presumably that desire supports my desire to be successful at more basic desires. How does that work?
I hope that the next hundred years of functional neuroimaging, incredibly detailed genetics and whatever else gives us much more knowledge of the brain that involves memories and experience more than what we must infer even today. I wonder if people will know by then what a desire is, why some feel a desire more than others, and how memories, knowledge, and doubt fit together, both functionally and biologically. If so, I think I'd be happier thinking about such things then.
One of my favourite song lyrics is from the Tragically Hip's At The Hundredth Meridian: "It would seem to me I remember every single f***ing thing I know."
I see the statistical difference. So what do these results really mean? How do those words differ then?
You're not presenting the most important aspect of all - how exactly did the participants interpret the know/remember distinction and whether it was the same for all or varied...?
I checked the article. Actually it was the researchers who specified the know/remember difference and they did it reasonably well: Remembering is when you have a specific recollection of when and how you heard the word; Knowing is when you only have a feeling you heard it, but cannot say when or in what context.
Another note on the results. They are quite interesting, but the most interesting part to me is that the overall correct rates are almost identical across both sets of experiments - with and without distruction at encoding stage. Therefore the difference in results is purely due to movement of some Remember cases into the Know category. Basically, the effect is trivial - encoding compromised, so you commit the word to memory less, so upon retrieval you cannot retrieve the full set of associations. Activation is weak, so you only have a "feeling". But cute none-the-less. :)
Mikhail's interpretation would also explain why things "known" about the past and "known" about the future "feel" the same. Both are derived from incomplete associations, one due to historical accident, the other out of necessity.
Interesting experiment. Let me comment, though, on the linguistics of "know" vs. "remember". The terms could be construed as English "evidentials". These are constructions that indicate the source of information. I could say the following:
"I know X is true" = logical certainty
"I think X is true" = uncertainty
"I heard/read X is true" = somewhat uncertain
"They say X is true" = conventional wisdom
Even though the researchers specified particular interpretations of "know" and "remember", the participants were probably biased by their normal use of these words, which may not have corresponded to the researchers' goals.
As a note, some languages have elaborate sets of grammatical evidential markers that make a wide variety of distinctions about how someone "knows" something.
I actually encountered this while perusing some medical education research based on Bloom's Taxonomy (http://www.sciencemag.org/cgi/reprint/319/5862/414.pdf). The medical researchers rate test questions according to Bloom's Taxonomy and then analyze the data using Bloom based ratings as ordinal data. But if recall and analysis as Gardiner and team report are not on the same continuum, then the analysis is flawed (the data should be categorical, at best). In other words, the study reported here confirms again that Bloom's Taxonomy is neither a hierarchy nor a chronology, though often misused as one or the other or both.