Uncategorized 08 Aug 2004 01:44 pm

CPU Architecture and Use of Natural Language

Ok, the following seems obvious now that I have put it into words, but I think it is fun and rewarding to think about anyway.

Recently, I read some advice on speaking and writing in the English language, and the gem that I took away was that it was best to use the smallest words possible to convey your meaning. The idea is that simplicity is best for communicating most effectively and to the largest audience.

As a child that grew up being drilled on ever more complex vocabulary at school, I have an inclination towards attempting to find the precise word for a meaning. That means that the arsenal includes a fair number of large words. The meanings for similar words are still subtly different.

I am not sure the subtlety gets across most of the time. There are some I meet who seem to pattern match against the gist of the conversation. To them the nuance or deviation from a simple word is lost, which means my meaning is lost. So, I started to think that the advice had even more weight.

Here’s another datum that confuses the issue. In computer architectures there were two schools of thought about cpu instructions sets. The CISC (complex instruction set) group believed that there should be an ever larger number of instructions available that do ever more complex things in the cpu. This makes it easier to program against with less code. The RISC school says that the number of hard coded instructions in the CPU should be small, and that longer combinations of those shorter instructions can produce the same effect as the single complex instruction. Actually, they claim that it will compute faster this way and be cheaper to make. In practice, today, most chips are a blend of both styles.

So, if we look at those instruction sets as our vocabulary, then maybe we should use small words in some situations and large words in others instead of always seeking to use small words.

Of course, making things simple is always harder to do. Guy Steele gave a great talk about building languages by starting with a very simple language and extended it by defining new, complex words in terms of the simple words. It is a very entertaining paper. So, it is probably a lazy inclination to use bigger words because they express more compactly, but you will be losing some percentage of your audience. Now in academic writing, which you find a lot of at the Arts and Letters Daily there is a high percentage of big words. It is interesting to contemplate how much of that is the need for precise meaning and how much of it is for unconsciously losing a percentage of the audience. Most of the time I think the goal is earnest precision. Sometimes, it seems like nonsense and snobbery is the goal. Definitely that causes some of us in the audience to walk out without the intended message.

Like I said, the observation seemed obvious once I made it, but the ramifications are still huge. There are lots of thought-provoking comparisons to be made that could enrich us everyday.

Here’s a fairly broad example. If you think of any human endeavor, there are some that are universal and some that are local. By universal I mean that the appeal is timeless and widespread. Being able to produce a universal work is a higher good than a local work becuase it touches more people. A key property of the works that are universal is simplicity. Local works tend to be “messy” because they require all kinds of knowledge about a particular place and time. Similarly, complex vocabulary requires the listener to have a bunch of specific knowledge in their head whereas simple language requires less from the listener allowing more listeners to be involved.

The trick is to get your point across. Hopefully, it can be beautiful too. I think the world has too much stuff that just fulfills the local need, and it is wanting for more of the universal stuff. Or maybe it is just me tired of reading drivel and wanting to be inspired by great reflection.

Subscribe to the comments through RSS Feed

Leave a Reply