Synthetic Intelligence

Prelude

First and foremost: please excuse the rambling nature of this post. If implicit (subconscious) memory is something that tickles your fancy, then I promise you’ll learn something if you sit through it. You’d probably learn something even if the subject matter didn’t tickle your fancy, but I’m just not sure I’d be willing to read through mad ramblings in something I’m not interested in. You know, unless it was written by me.

Outline

1. Basic information regarding explicit memory
2. Basic information regarding implicit memory
3. Methods of testing memory
4. More in-depth information on implicit memory
5. Implicit memory illusions and biases

Basic information regarding explicit memory

Explicit memory is the word used to describe a person’s conscious memory. Every memory that you are aware of is an explicit memory; eating dinner last night, showering this morning, your first birthday party, the last day of school—all of it.

The above examples are all episodic memory, which is a form of explicit memory. Additionally, semantic memory fits the bill here: words and facts and the things you’ve consciously learned are also a form of explicit memory.

Basic information regarding implicit memory

Implicit memory is commonly referred to as unconscious memory or indirect memory. It’s automatic memory; you’re not even aware that you know the things you do, a lot of the time. The most common type of implicit memory you have is procedural memory—the knowledge of how to do things. If you know how to swim, that’s procedural memory. If you know how to walk, or to run, or to jump, or to ride a bike, drive a car, draw a circle—the memories in your head that tell you (automatically, in most cases) how to do these things are part of your implicit memory.

Methods of testing memory

Implicit and explicit memory are tested in very different ways.

Explicit memory, due to the inherent nature of having conscious knowledge of the memories you’re storing, is fairly straight-forward to test. The easiest method is simply standard recognition: Have you seen this face before? Have you seen this house before? Have you ever been to this park? The other major method of measuring explicit memory is to have study participants recall the information they’ve stored: What color shirt was the clerk wearing? How old are you? When did you graduate from high school?

Implicit memory, on the other hand, is harder to measure. And for good reason, too—you don’t know what you know. Many different methods are exercised to invoke tangible measurements.

1. One of which is tachistoscopic recognition, where someone is shown an image for a fraction of a second and tested on what can be remembered.
2. Priming is another process where people will demonstrate improved performance on tasks where they have been previously prepared for subconsciously.
3. A third common method for testing implicit memory is with word-stem completion. Assume I’ve been secretly flashing images of breakfast foods for fractions of a second on your screen while you’ve been reading this. You didn’t see it; you don’t even know that I did it until I just told you. But then if I introduce you to a word you need to guess and fill in a few letters at the beginning (C e _ _ _ _, or E _ _ _), you’re going to be more likely to answer with something breakfast-related because it’s fresh on your mind—whether you know it or not.

In both of these types of memory, one fact remains true: the memory’s connection strength can be measured by response time. Things that you’ve recently seen, thought about, or heard come to mind faster than others.

More in-depth information on implicit memory

I’m sure you’ve all experienced the phenomenon known as “Deja vu,” where you have a distinct feeling of familiarity but you don’t know why. The effect stems from your implicit memory. Remember that for a minute while you read on—it’ll all make sense soon enough.

There is so much visual stimuli in the world that your brain couldn’t handle it all if it tried. Instead, your eyes act as a funnel to focus on the most important things. The things that you are consciously focusing on are processed in your working memory (WM) and then stored into long-term memory if needed. Implicit memories, on the other hand, bypass your WM and do not require conscious processing to store or retrieve. In other words, you’re subconsciously processing and storing a lot of what you’re seeing, but not focusing on.

So, when you get deja vu when you see that guy at the supermarket that you know you’ve seen before, it’s very possible that you have seen him before—just that you don’t know you did.

On a somewhat unrelated subject, explicit memory relies on the hippocampus (where WM is processed) for consolidating and encoding memories; implicit memory doesn’t. We know this because of a famous case of a guy named H.M., who suffered from anterograde amnesia, preventing him from forming any new memories. During his course of amnesia, his mother died. Afterwards, he would ask about her daily, and every day would be told about her death, causing him to cry and grieve. He never learned that his mother had passed away, but instead began to associate asking about her with the pain he felt when told she had died. Eventually he stopped asking.

Another example, dealing with amnesia patients, is in procedural memory. Those who cannot create new explicit memories have a slight difficulty learning new skills, such as how to ride a bike or paint, but can learn them. Once learned, these people can easily repeat the performance at a later time as if they knew how to do them already.

Procedural memory is interesting: it starts out as a conscious effort to learn—explicit memory—but eventually becomes automatic: implicit memory. Once it’s switched to implicit memory, you no longer have to think about doing it—you can just do it. This is the case with driving: anyone who has been driving a while doesn’t have to concern themselves with what gear they’re in or when to turn on their blinkers—they can instead focus on driving safely.

Implicit memory illusions and biases

• Illusion of Fame (Jacoby et al., 1989)

  In an experiment, a group of participants were presented three groups of names:

  1. Famous names
  2. Somewhat famous names (famous people that weren’t that popular)
  3. Made-up names

  Later, they were asked to recall the names and how famous they are.

  The results found that subjects forgot where they had seen the names before, and consistently rated most of the made-up names as famous as others. They’d forgotten the source of the information, and the psychologists concluded that familiarity lasts longer than the source memory.

• Mere Exposure Effect

  Being exposed to something multiple times makes you like that thing more, even if you aren’t able to recall the memory of seeing that thing later. In other words, if you want to get a job somewhere, pop in once or twice a week for a couple weeks prior and make sure the person in charge of hiring sees you.

• Illusion of Truth (Begg et al., 1992)

  Repeated messages are more likely to be accepted as true, even if you say it is false as you say the message.

• Source Confusion (Brown et al., 1977)

  Familiar faces or faces that someone has seen before (been primed with) are more likely to be picked out in a line-up.

• Stereotype Susceptibility (Shih et al., 1999)

  A large group of Asian women were split up into three salients focusing on different traits. When preparing for the experiment, they were given three different questionaires to gather information about them:

  1. The female salient was asked questions regarding their gender, their habits, and whether they live in co-ed/single-sex dorms. This questionaire got them primed on their “female-ness”.
  2. The Asian salient was asked questions about their culture, the languages they speak, and their family. This questionaire got them priced on their “Asian-ness”.
  3. And then there was the control salient that was asked general questions regarding their name, age, location, and so on.

  And then the participants were given a math test. Group 2 (the Asian-primed salient) performed the best. Group 1 (the female-primed salient) performed the worst. Stereotypes, for the win!

• Other studies

  • Study of the Elderly (Levy, 1996)
  • Study of African Americans (Steele & Aronson, 1995)