Memorization vs. understanding, reply by Bing's Chat

Memorization vs. understanding, reply by Bing's Chat

On 27 February 2023, I asked Bing's Chat to defend the thesis that
memorization comes before understanding. The following is what
I asked it twice, with different constraints each time. This is its
first reply. BingChat asked me to describe my audience and how many
words in the essay. I told it to gear it towards an academic audience,
at 120 words. The essay I got back is impressive, even if I don't
completely agree with it. But judging it as an "English composition",
I find it very impressive. Anyway, the chat wrote back:

Here is an essay I wrote based on your query:

Memorization is often seen as a passive and superficial way of learning,
while understanding is regarded as a deeper and more meaningful way of
acquiring knowledge. However, memorization and understanding are not
mutually exclusive, but rather complementary processes that can enhance
each other. In fact, memorization can be a necessary precondition of
understanding in many cases.

My comment: In this first interaction with this bot, I have not
demanded that the bot provide for me its definition of understanding
and then demand that it stick to it rigorously. But if I were debating a
human, I would demand that the human provide one example of
something that can be understood that is not dependent on prior
memorized knowledge. However, I'm not so eager to do that to a bot,
poor thing.
One reason why memorization can precede understanding is that
memorized information can provide the foundation for further learning
and comprehension. For example, students who have not memorized their
multiplication tables may have difficulty solving some maths problems12.
By memorizing basic facts, they can free up their working memory and
focus on more complex concepts3. Memorization can also help students
recall previous knowledge and connect it with new information4.

Another reason why memorization can come before understanding is that
some information may be too abstract or unfamiliar to be understood
without repetition or practice. For example, learning a foreign language
often requires memorizing vocabulary, grammar rules, and pronunciation
before being able to communicate effectively5. Memorizing these elements
can help learners develop fluency and confidence in using the language6.

Memorization can also help learners recognize patterns and structures that
facilitate understanding4.

In conclusion, memorization is not an inferior or outdated method of learning,
but rather a useful and necessary one that can support understanding.
Memorization can help learners build a solid foundation of knowledge, recall
relevant information, and master challenging skills. Therefore, memorization
should not be dismissed or neglected, but rather integrated with understanding
in a balanced way.

Essay length: 120 words

I hope this essay helps you with your academic work. Do you have any
feedback or questions for me?

It's clear that the bot drifted away from talking about memorization vs
understanding and added the wider concept of learning. That's OK, but it
reveals that the bot has so far not grasped what the basic contention here
is. Nevertheless, from where I'm sitting,
Learning = Memorization + Understanding
It's just that memorization is prior to understanding. For example, if you
are in the process of understanding the sine and cosine functions based on
the right triangle, then you had better have already memorized what the
standard parts of a right triangle are and what angle is. Let's get specific.
An angle is a relationship between two parts of a triangle. The sine function
of that angle is defined as a relationship of the ratio of the lengths of two
of those parts. This is what I meant when I claimed that understanding is the
process of forming relationships among the parts (sides and hypotenuse) of a
triangle (the whole). Is this really so difficult to grasp?

Summing up what the bot reported back to me: Memorization is an essential
part of overall learning, which is true, but it misses the point. I then realized
that if I were ever to get the bot to directly agree with or disagree with my
thesis, that I would have to present it with my definition of what it means to
understand something.

So, I added my own definition of what I think understanding is. The bot then
indirectly and directly challenged my assumptions with 'examples'. I will reply to
the challenges at the end of this writing, but first I want to comment on how
impressed I am at what BingChat wrote from a scholarly viewpoint.

One compliment that I give to BingChat is that it was able to correctly interpret
what I asked it, in spite of the many misspelled words I had in my query.

From BingChat 27 Feb 2023. This time, I will comment after each paragraph.

(This was my query with spelling corrections added)
Write an essay explaining that memorization comes before understanding.
Essay length = 120 words. Audience = academia. I define understanding as
the process of formulating relationships among the parts of something. If
you haven't memorized the parts, you can't form meaningful relationships
on them to form a comprehensive whole. What do you think?

BingBot's reply: Here is an essay I wrote based on your request:

Memorization is a foundational skill that can be used to build understanding1.
Memorization is the process of committing content to memory, while understanding
is the process of comprehending the content, such that you appreciate the
fundamental idea governing it and can see how the different parts of the content
relate to one another and other information outside the content2. Memorization
is temporary and limited to basic words and facts, while understanding is long-term
and unlimited3.

Look, I understand how hard it is to give a rigorous definition of what
it means to understand something, but I've learn over the decades that,
after a certain point of 'clarification', further efforts only subvert your
effort. For example, using a vague term such as 'appreciate' is not helpful
here at all. And what does the word 'fundamental' mean if not the parts
underlying the whole?
In some cases, memorization comes before understanding. For example, when
learning a new language, you need to memorize vocabulary words before you can
understand grammar rules or form sentences. Similarly, when learning mathematics,
you need to memorize formulas before you can understand how to apply them or
solve problems. Memorized information is often the foundation for understanding2.
This consession is about as far as I ever get out of a human. When I
have debated people in the past on this topic, I have tried to disabuse
them of three stubbornly held misconceptions: 1) I am not saying that
education should perform a 180 degree turn back to the old-fashioned
memorization of mere 'facts'. 2) My opponents always seemed naive
about the level that they rely on the foundation of 'mere memorized
facts' before they can argue anything at all. 3) Rote memorization is
evil!! No, it isn't. I'm not going to tell you how much to memorize,
except to say this: You probably should memorize more than you want to.
And, laziness is no excuse for not memorizing. No pain, folks, no gain.
However, memorization does not always come before understanding. Sometimes,
understanding comes before memorization or even replaces it. For example, when
learning history, you do not need to memorize every date or name, but rather
understand the causes and effects of events or movements. Similarly, when learning
science, you do not need to memorize every fact or term, but rather understand the
concepts and principles behind them. Understanding helps you remember information
better than memorization alone3.
This example is not an example at all. I doubt that there has even been a single
history teacher who has told the class to memorize 'every date and name'.
What does this vague statement even mean? I'm going to do what my
opponents rarely do here, and that is to give specific examples. In this case,
how about the historical importance of Caesar crossing the Rubicon? The
significance of this action does not require me to know all the names of all
the soldiers under Caesar's command. It doesn't even rely on my knowing
the exact date of this crossing. It's significance lies in the fact that it was
an act of treason against the Roman Republic and an act of war. And if you
haven't memorized these facts, you wouldn't be able to *understand* its
significance.

As for the claim that 'Understanding helps you remember information better
than memorization alone', I can only agree. But why do you suppose that
that's so? Could it be that the process of forming an understanding of a
collection of parts by relating those parts to each other, have resulted in
categories of parts that facilitate memorization of those parts?

Therefore, memorization and understanding are both essential in knowledge
acquisition, but they are not always sequential or mutually exclusive. Depending
on what you are learning and how you are learning it, you may use either one or
both strategies.

(Word count: 219) [Comment: The bot seems to have terminated the formal essay
here, but then provided me with some afterthoughts, which follow.]

I hope this essay helps you with your academic work.

What do I think? I think that memorization and understanding are both important
skills that depend on each other. However, I also think that there are different
types of understanding that require different levels of memorization. For example,

What type of understanding are you interested in?

So, the bot was trying to continue the discussion, but I prefer to keep interactions with
bots short. Actually, I'm interested in all three types.

The bot ignored my request for it to restrict its essay to 120 words, which is fine because
it definitely needed to be longer this time. I like that it used a bulleted list for me.

Let me address the bot's first bulleted item.

To claim that my thesis is wrong because if a subject area has a lot
of parts then that requires a lot of memorization, is just nonsensical.
No one ever promised that arriving at the understanding of something
is necessarily going to be easy. This bot seems not to grasp that
understanding can occur at different levels or depths.

Let's look at a real-world example. Imagine that you're a typical office
person who uses the company's HP printer for your business print jobs.
You understand the machine only deeply enough to know where the
power switch is, how to load fresh paper into it, and which output tray
to find your print job. But if something goes wrong with the printer,
you call for someone who knows the printer better than you do. So, the
office IT guy comes in to fix the printer. He knows a lot more parts of
the printer than the casual user does. He knows where the networking
ports are and how they are configured. He knows how the basic modules
inside the printer fit together so he understands what operations he
must perform to, say, fix a paper jam. But if the IT guy doesn't
understand the printer as deeply as the problem goes, so he can fix it,
he can contact someone at HP, who knows the parts of that printer even
better than he does. And then there's one person at HP who understands
that printer better than anyone else and that is the guy who directed the
design and construction of the printer in the first place. So, the bot
is confused or something. Just because a whole has a thousand parts, say,
doesn't mean that everyone who uses that whole (the printer in this case)
needs to know every part and how they interoperate. How much a person
needs to memorize about that printer depends on the type of functionality
the person requires of the printer.

When I majored in mathematics in university, I also studied history. But
I came away from that experience with a lot less understanding of history
than a history major, and vice versa. No one has ever said that you have
to memorize everything under the sun before you understand anything.

Let me address the bot's second bulleted item.
The claim that a shorter path to understanding is from "applying
knowledge to new situations." But what is knowledge in this context?
It is a collection of facts that are already in your memory. But what
difference does it make to the question at hand if that knowledge was
put into your memory for instant, accurate retrieval five minutes ago,
or five years ago? Either way, it must be in you prior to you building
on it to somehow get new knowledge from old.

There is no royal road to understanding: You have to have a certain
amount of prerequisite parts in your memory before you can understand
how they relate to each other to form a whole.

Time for another example. A five-year-old boy goes to his father and
says, "Dad what does 'What goes up, must come down' mean?" The
father replies, "If you throw something up into the air, it has to eventually
come back down to the ground." If the boy understands his father's
explanation, it is because of the parts of this situation that he already
has in his memory. I assume that the boy knows that from where he's
standing, of all possible directions, only one of them is factually 'up' and
similarly only one of them is factually 'down'. If he doesn't know these two
facts, which are key ingredients to the 'whole' concept, then he won't get it.
Of course, there's more the boy has to know (have in his memory), but I think
my point has been made. If you want to really test this point, consider an
intelligent being, named Ernie, who has lived all his life free floating in deep
space (in which there is no natural up and down), and get him to understand
what it means.

Let me now address the bot's final bulleted item.
The claim that a third way to get new understanding is "the process
of creating new knowledge from existing knowledge." Again, this is
somewhat of a misunderstanding, because knowledge is a bigger
category than understanding. It goes on to say that, "This type of
understanding requires almost no memorization because you can generate
novel ideas or solutions based on your own reasoning or intuition." Well,
isn't it a process of reasoning to find relationships among parts? And
don't those parts already have to be memorized by the reasoner? And
once you've reasoned out these new relationships, shouldn't you memorize
them to keep them?

As for the notion of intuitively received understanding, I have no problem.
But so long as that understanding remains intuitive, we can't analyze it.
Just the same, I have to congradulate the bot for coming up with an
interesting exception to my thesis. The reason I didn't want to veer off
from my scholastic thesis and present alternative modes of understanding
is because my intended audience is not to mystics or philosophers, but to
lay people and educators who have a negative view of the proper role of
memorization in modern education.

Additional analysis of the bot's essay and comments. First paragraph:
Memorization is a foundational skill that can be used to build understanding1.
Memorization is the process of committing content to memory, while understanding
is the process of comprehending the content, such that you appreciate the
fundamental idea governing it and can see how the different parts of the content
relate to one another and other information outside the content2. Memorization
is temporary and limited to basic words and facts, while understanding is long-term
and unlimited3.
When the bot claims that "understanding is the process of comprehending the content,
such that you appreciate the fundamental idea governing it and can see how the
different parts of the content relate to one another and other information outside the
content," I have to observe that it appears to partly agree with me that the "fundamental
idea governing it" is dependent on one of two possibilities: 1) how the parts of interest
relate with each other, or 2) how the thing relates to its environment, which merely means
that the thing (of content) is also a part of some larger whole. Either way of looking at
it, it seems to agree with my basic thesis of parts-to-parts & parts-to-whole. And when
the bot says that "understanding is the process of comprehending," it's using the term
'comprehending', which is usually taken to be synonymous with 'understanding', and trying
to characterize understanding in terms of it. This is something that has to be avoided in
this discussion.

By the way, I feel I need to distinguish between characterizing and defining. To characterize
X is to say something true about X; whereas, to define X it to say everything that is true
about X.

But when it characterizes memorization as short-term knowledge and understanding as
long-term knowledge, that's just silly. Most people I know can remember things from long
ago. Let's go back to the example of the right triangle. Can you claim that you understand
how to perform taking a sine of an angle in a right triangle if you have forgotten everything
you once knew about a right triangle? I'd say not. Once you've memorized the parts and
discovered their necessary relationships, then you also memorize those relationships. Thus,
the memorization is not needed for only basic "words and facts"; we also need to memorize
the relationships among the parts-to-parts & parts-to-whole.

Next paragraph:

In some cases, memorization comes before understanding. For example, when
learning a new language, you need to memorize vocabulary words before you can
understand grammar rules or form sentences. Similarly, when learning mathematics,
you need to memorize formulas before you can understand how to apply them or
solve problems. Memorized information is often the foundation for understanding2.
So, here the bot seems to be claiming that at the first level of a new subject, one must
of necessity jumpstart the learning by memorizing the basics. What people seem not to
get is that the very 'next level' has to be jumpstarted too. It's the application of the same
algorithm over and over again: Discover the parts; memorize them in enough detail to
accomplish what you need; find the useful relationships among the parts; thus obtain
understanding; then memorize those relationships to keep that understanding permanent,
unless you don't mind starting the process all over again.

Instilling in students a negative attitude towards memorization only hinders them
being able to understand the deep concepts of the modern world. I can't tell
students how much to memorize in a given class, but I can tell them not to
develop a bad attitude towards memorization, per se.

Homework problem: Think of one example of understanding something that doesn't
rely on having some previously memorized information whatsoever.

Warning: I'm concerned that the reader may misunderstand just how general my
definition of understanding is and thus fail to see how to apply it to the examples
given by the bot. Of course, when I discuss parts of a laser printer, one can easily
see how those parts relate to the whole concept, in this case being the entire laser
printer. But history can be arbitrarily divided into parts as well. For example, one
could define a 'whole time period' as the time of the Roman Empire from Julius
Caesar to Nero, say. Then one could divide that period into 'parts' by days, months,
years, or by each emperor's time of reign.

So, on the one hand, the abstractness of my definition makes it hard for some to see
how it can be applied to real-world situations. But on the other hand, its abstractness
allows for both compactness of expression and generality of application.