Wednesday, July 22, 2009

Tomato, Tomato

I speak english (or more precisely, the Queen's english) with a slight tinge of Geordie when I get angry. I sometimes say things that other english speakers fail to understand - either I pronounce words in a way they don't understand, or I use words that I think mean a certain thing when my conversational partner means something else.

It never fails to amaze me that we actually manage to communicate ideas and concepts to each other, given the opportunities for vagueness and overloaded meanings of words, both spoken and written. Lead, lead - one's a metal, one's where I guide someone. Edinburgh - but don't say "burg".

At my workplace, we have a set of words that are massively overloaded. Build - what does that mean to you? for me, depending on who I'm talking to and the time of day, it could mean building some code, building some data, the result of a lighting calculation, the latest version of the game, and a variety of other potential options. Pipeline and workflow suffer the same issues, and don't get me started on "Object", "Entity", "Component" or any of the other synonyms for "Thing".

The worst example I saw recently was an MIT video (Introduction to algorithms) where the presenter delves into formal proofs for complexity representation, and midway through one of his equations, makes a quick aside to the effect of "That equals sign - well, that really means definitely not larger than in this context". I have 30-something years of brain mapping that says the symbol "=" means "equal to", and he's using it to mean something different. Surely there's a better symbol for him to use? maybe even "<=" ?

This overloading and confusion crops up constantly, and it's a huge stumbling block whenever people need to collaborate. Sure, some of these issues could be resolved by being more precise when talking or writing, or establishing glossaries when communicating. However, I think the issue is much deeper than that.

Our brains are heavily wired into attaching a meaning (or a feeling) to symbols. Words, mathematical symbols, brand images, even blocks of raw colour evoke a response. Pictures really are worth a thousand words, and look just how important words are. Misusing a word or using a word in a different context is a problem when your brain is trained to respond to that word (or sound, or picture) with a preset response.

Another example that came up during my discussion with Joe was the symbols we use in calculus. When learning to differentiate at school, I had endless problems - not with the concepts, but with the notation. every time I saw "dy/dx" my brain was screaming at me - Divide! Divide you moron! You've done it a thousand times, what makes this time different? This problem never really went away until I saw Newton's notation (dots above letters; more dots equates to higher order derivatives) and Lagrange's notation. Seeing a way of expressing the concept where the very expression itself wasn't already mentally mapped to something else instantly removed that roadblock to how I thought about the process.

The same issues are also prevalent in UI design. Pictures on buttons - or if you drive an american car, english words on buttons - become ingrained from the moment you see them, and if you see a different image mapped to a familiar operation, or the same image meaning something else in a different program - you are instantly confused.

So, next time you strike up a conversation with someone, don't just say what you want to say - say what you want them to understand. If that makes any sense at all. Either you'll confuse the hell out of yourself, or you might just get your point across.

Did I mention that my tomato plant is doing really nicely?


  1. All this stuff has been well researched already in the field of knowledge management. If you want a quick tour then I'd suggest as a good starting point - particularly for anybody that's done a spot of sampling. A detour to is then recommended before reaching the final destination of where it becomes clear that language shapes our learning process, and hence how we understand the world around us. The journey is then complete in terms of understanding why Newspeak (or NuSpeak - the language of NuLab) is so important - control the message and you control the meaning.

  2. Excellent links, Chris. Given the depths of research, does it not make you wonder that the areas where communication modification (propaganda, misdirection) appear most prevalent seem to be advertising and politics? Why do the general public not see this happening around them? We're being lied to more effectively than we're being given the truth.

    Taking the tin foil hat off, How can we apply our symbol alphabets to the task of communicating with less error? Assuming a larger alphabet (like Kanji) or more error correction (I haven't got a good example language here, but recent computer programming languages tend in this direction via syntax) is there a measurable property that describes learning speed or context transmission to point the way?

  3. to give some background context, this post came on the back of Amazon redacting 1984, which you have to wonder is a publicity stunt to point out the absurdity of sensorship (given the premise of information reduction as a form of control). Samuel R. Delany's Babel 17 pushes the other way, and suggests complete clarity of communication given the perfect language. I wonder if the mix of media we are becoming more accustomed to with the interwebs (language as text plus sound plus colour plus ...) will change how we communicate.