Find here a rough summary of Chapter 3 of The Design of Everyday Things by Don Norman along with my attempt at critical responses to some key passages.
Chapter 3
First, a summary of key points that Norman lays out at the beginning of the chapter:
- Knowledge is both in the head (memory) and in the world (to be perceived and interpreted)
- Great precision of knowledge (remembered, retrieved, or perceived) is not required for precision of behavior
- Natural constraints exist in the world that limit possible actions and interpretations
- Cultural constrains and conventions exist, even if only in the head
He demonstrates this with an example:
“In knowing what our currency looks like, we don’t need to know all the details, simply sufficient knowledge to distinguish one value of currency from another… because so much knowledge is available in the environment, it is surprising how little we need to learn.”
-The Design of Everyday Things, p. 77
This makes me think of ways that I’ve heard the brain is efficiently designed. This particular quote outlines how the brain operates efficiently by artfully combining the storage and perception/interpretation mechanisms to optimize the overall power and effort expenditure. I also saw a video (ok, fine, a TikTok) of Jeff Bezos saying that AI models require power on the order of many kilowatts to do the same things that human brains can do with just about 20 watts of power (@lexfridman, TikTok 2023). In this chapter, Norman also discusses how retrieving details from long-term memory (try to remember what you ate for lunch two Saturdays ago) takes significantly more effort than retrieval from short-term/working memory (try to remember what I just asked you to retrieve from your long-term memory). I tried to find an estimate of the power required for each, but couldn’t find anything. I did find, however, that according to Scientific American the brain has a storage capacity estimated to be around 2.5 petabytes (Reber, 2010). For reference, there are a million gigabytes in a petabyte and a normal laptop might have 256 gigabytes of storage.
Norman then makes a key distinction between two types of knowledge: declarative knowledge (easy to write down, facts) vs. procedural knowledge (hard to write down, how to actually do things). According to a paper I found, the systems responsible for each of these memory systems is fairly distinct, but sometimes the systems overlap in responsibilities (namely with grammar rules and navigation) (Lum, A G et al). From what I could understand in the paper, it seems like the parts of the brain that support declarative knowledge include memory and learning functions, while the procedural knowledge is, as you might expect, supported by parts of the brain with more diverse functions such as some related to motor control.
Norman then starts to discuss how we’ve designed our world to enable even more efficiency in the brain:
“… Strong constraints… simplify what must be retained in memory”
-The Design of Everyday Things, p. 85
The benefit of constraints is that they limit what is possible, thus reducing the need for as many details to be retained (or retrieved) in order to figure out an answer. This makes me think of a fun word trick I saw years ago in which letters in words can be scrambled in each word of a sentence except for the first and last letters, and readers will find that the sentence is still completely readable. Example: I bet you can tllatoy sltil raed tihs wolhe stnecnee eevn wtih the wdros tihs srbmlecad. This exemplifies a phenomenon similar to the one in the quote, implying that even if somewhere down deep we know the order of letters in words, we don’t need to retrieve the exact order of letters but rather just the key letters and length in order to distinguish the word from others given appropriate context. The letters listed, length of the word, and context in a sentence prove to be enough constraints to limit the possibility of what word it is such that it isn’t essential to match the exact order of the letters. My brother, who studies psychology, says this is a prime example of what is called “top-down processing.”
Norman then discusses that the more recent move that many companies have made to require stronger, unique passwords to boost security actually forces people to rely on relatively insecure password storage mechanisms because they can’t remember all of the unique, complex passwords with their brain alone. He also talks about the fact that security theory outlines that the best identity authentication relies on both “‘something you have’ plus ‘something you know’” (Norman, p. 91). Since the book has come out there has been a proliferation of two-factor authentication methods for verifying identify, which would be categorized under “something you have.” Despite this growth, it doesn’t seem like the complexity requirements of the “thing you know” have been reduced in response to two-factor authentication, and the risk of insecurely stored but sufficiently complex passwords persists. In other words, the designers of security systems are still consistently over-estimating the abilities of human memory.
After covering the basics of memory, Norman outlines a few general tips for designers:
- Don’t expect people to memorize things you’ve just shown them. The example that comes to mind is verification codes: to verify your identity during log in, a website or app will send your phone, email, or device a 3-7 digit code that you enter into the site or app to complete verification. From personal experience I find 7 digits to be the point at which I feel like I’m pushing the limits of my short-term memory. I suppose these companies have had to weigh the difficulty (aka number of digits) for a hacker to guess the correct code with the feasibility of the user memorizing the code and landed on 6 or 7.
- Sensory memories don’t interfere across senses, so it will be easier for a user to remember more information if it is delivered across multiple senses. I’m thinking this could be most applicable to marketing and advertising, but the only example I can think of is a dessert store or bakery that has music, signs and decorations, and the smell of the food wafting around. The other application I can think of is education, but again this doesn’t feel super relevant.
This chapter ends with Norman’s discussion of natural mappings. He explains that a lot of our mappings are illogical, but if they are relatively consistent in the tools and contexts that we use, then they can still be effective. The main issues that come up are when mappings are both illogical and inconsistent (see my microwave rant in the previous post). Another painful-great example that comes to mind are the handles on sinks, where the main issue is how inconsistent they are across systems. Particularly interesting was Norman’s explanation of how natural mappings that define how we think of things like up and down, forward and backward, can vary based on cultural interpretations of the core meaning of these things.
And finally, a quote to end on:
“The unaided mind is surprisingly limited. It is things that make us smart.”
-The Design of Everyday Things, p. 104
Works Cited
Fridman, Lex @lexfridman. “Jeff Bezos on why humans are special – clip from Lex Fridman Podcast #405 with Jeff Bezos. Guest bio: Jeff Bezos is the founder of Amazon and Blue Origin.” TikTok. December 2023.
Lum, Jarrad A G et al. “Working, declarative and procedural memory in specific language impairment.” Cortex; a journal devoted to the study of the nervous system and behavior vol. 48,9 (2012): 1138-54. doi:10.1016/j.cortex.2011.06.001
Norman, Donald A. The Design of Everyday Things. MIT Press, 2013.
Reber, Paul. “What is the Memory Capacity of the Human Brain?” Scientific American. Web. 1 May 2010. https://www.scientificamerican.com/article/what-is-the-memory-capacity/.
BYOMSPM 
Leave a comment