This is a question inspired by this recent question on the Chinese Language & Usage website.

Someone asked why they needed to learn how to write Chinese characters, since today we mostly use computers or text messages to do so. Several users replied that writing Chinese characters help remembering them more easily, including for reading purposes.

However, another user argued that there was no reason to believe that it was true, and that writing characters doesn't necessarily help one remember them (compared for example to simply reading them and looking at them).

Is that the case? Does writing generally help remember glyphs (or more generally information) or not? Is there any conclusive research about this or about the relation between writing and memorizing in general?

  • 4
    $\begingroup$ From a high level neurological perspective (I'm no expert, hence the comment), I would expect any extra links to what you want to memorize help in memorizing. $\endgroup$
    – Steven Jeuris
    Commented Jan 19, 2012 at 13:51
  • $\begingroup$ I have noticed that writing out theorems in Euclid's geometry have helped me remember them better in my high school.. so much so that i have scored well just by recapping in an hour.. I think classical psychologists will call it procedural memory. $\endgroup$ Commented Jan 19, 2012 at 14:24
  • $\begingroup$ This is a phenomenon applicable generally across all disciplines. Chinese language or any other cognitive process has any particular relevance. It has been observed in day to day activities that writing down something is one way of strengthening its impression in memory. It is only to be expected that this will be corroborated by scientific studies. I may post an answer if I compile suitable material. $\endgroup$
    – Kris
    Commented Jan 19, 2012 at 15:18
  • 4
    $\begingroup$ Here's a related question from Skeptics Exchange: skeptics.stackexchange.com/questions/5818/… $\endgroup$ Commented Jan 20, 2012 at 4:00
  • $\begingroup$ @Solus: I checked out skepticsSE. The post remains inconclusive it seems to me, in spite of your excellent elaborate answer there. $\endgroup$
    – Kris
    Commented Jan 23, 2012 at 6:08

3 Answers 3


Yes, writing increases the modality and attention given to a piece of information. Increasing the effort and the ways that you have experienced a bit of information helps you encode that information better; this is Elaborative Encoding.

More generally the more deeply you process a thing the more likely you are to properly encode the memory for future retrieval. Simply looking at characters is very shallow processing; writing and constructing them is relatively deep. This is the Modality Effect.

Explaining how things work and showing others or doing a task is also a much better way to remember things than just reading about them; in the case of remembering Chinese Characters there's less you can do other than write them, but practice methods like writing a brief paper can help you learn to use them. This is why essays are a common task in language classes.

Note however that Transfer-Appropriate Processing is also relevant: retrieval is more likely when we attempt to retrieve information in a similar or appropriate context to when encoding occurred. Therefore writing symbols could arguably be a bit less appropriate way to learn to read symbols; however it is an extremely relevant way of learning how to write and construct those symbols.

I've found some general reference sites eluding to this effect such as this post on Lifehack: Writing and Remembering: Why We Remember What We Write, but could not find experimental results supporting this specific idea. I know my advanced course on Memory covered this specifically but I don't recall exact studies.

  • $\begingroup$ So, if we take the view that learning consists of building schematic connections, and that recall depends upon the number of connections we have, then it stands to reason here that the reason writing helps with retrieval is that it allows for the creation of more connections. This seems like a theory that could be evaluated empirically, not sure if that research has been done. $\endgroup$
    – theMayer
    Commented Feb 28, 2014 at 22:09
  • $\begingroup$ Great answer. At the same time, I think we are better at memorizing something when the stakes are high. After writing something down and storing it in a safe place, there is not much lost by forgetting the information anymore. $\endgroup$
    – danijar
    Commented Nov 1, 2020 at 21:03

I'm going to disagree with Ben here. My colleague Adam Putnam has spent several years researching whether it's best for memory to speak, write, or even think particular responses out loud. His research has continued to turn up no differences between these different modalities, despite what we know about transfer-appropriate processing and elaborative encoding. He has a paper in preparation on the topic, and I'll be sure to link it in should it be accepted.

Other research on the production effect, however, does suggest that "producing" a response should lead to better memory than just reading it. Theories suggest that producing a response increases the distinctiveness of that particular memory. Distinctiveness has less of an effect, though, when the same manipulation is applied to multiple items (e.g., the produced memory is less unique if surrounded by other produced memories). There are several theories out there that explain why this might be.

Our lab's research, though, on the bottom line, shows that it's not the way that you record your answer inasmuch as whether or not you're testing yourself as you do so. For instance, consider the Chinese character you want to learn, imagine it, and then check yourself -- that should be very similar to considering the character, writing it down, and then checking yourself.

  • 4
    $\begingroup$ Interesting answer; I think your answer would be more interesting with some citations. Any chance you could edit your answer to include a few? $\endgroup$ Commented Jan 30, 2012 at 4:43
  • 2
    $\begingroup$ On a slightly related note, I wrote a Python flashcard script for testing myself on the Periodic table. I typed, not wrote, and it certainly helped (at least seemed to for me). If I got the element (or name, depending on which way I was testing) wrong, it would be returned to the pile, until the "stack" was empty. That lends weight, for me, to your assertion that the importance is more on testing than simply the act of construction. $\endgroup$ Commented Feb 1, 2012 at 13:40
  • 1
    $\begingroup$ Wayne, you may be interested in a 2007 paper by Karpicke and Roediger, who speak to this exact issue. They found that it's even more beneficial to keep returning flashcards to the stack until you've retrieved each multiple times. You'll find a comment I wrote on another question relevant, too. $\endgroup$ Commented Feb 6, 2012 at 15:20

The behavioral output itself can only indirectly affect memory, but the mechanism of reinforcement lies in its reception of attention. For instance, athletes who practice a sport seriously know, as also popularly attested by Mack and Cassteven's Mind Gym, that 50% of practice should be mental. They should watch the sport, imagine playing the sport, think about how they want to improve, think about winning, etc. It doesn't matter if the sport is physically practiced or imagined, as long as there is attention.

The route that monotonous rehearsal plays in memory is in forcing attention onto the content, and it most likely is a better strategy than, say, watching TV; however, that doesn't mean that other strategies for enhancing memory aren't far more superior. For instance, as Ben said, the deeper something is encoded - embedded with meaning - the more likely it will be learned.

  • 1
    $\begingroup$ This is a nice, but it would be great if you could throw in some references. Also, welcome! $\endgroup$ Commented Aug 10, 2012 at 22:13

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

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