Engineered Languages

Chapter 3 looked at languages built to be spoken between people. An engineered language (engelang) is built for a different reason: to test an idea. The designer wants to know what happens to a language if one feature is pushed to an extreme. Is it possible to have a language with no ambiguity? How much can you say with a hundred words? This chapter follows two engelangs that answer those questions in opposite ways, then asks the big question behind both: does the language you speak shape the way you think?

Lojban: A Language Built on Logic

From Loglan to Lojban

Loglan was invented by James Cooke Brown, who began the work in 1955 and described it publicly in Scientific American in 1960. His aim was to examine whether language influences thought: if you gave people a language with a logical, unambiguous grammar, would it change how they reason? Disputes over the copyright of Loglan's words led a group of supporters to leave and found the Logical Language Group, who in 1986 decided to rebuild the entire vocabulary under a freely usable licence. The result was Lojban ("logical language"), with a baseline description completed in 1997 and published as The Complete Lojban Language.

How Lojban Works

The core of Lojban is the same idea as the logic of predicates. A sentence is built around a selbri (the relation) and its sumti (the things it relates). Every relation word comes with a fixed list of places, in a fixed order, so you always know what each slot means.

The vocabulary is also built to a plan. The root words, called gismu, number in the region of a thousand and three hundred and were chosen to be as recognisable as possible to speakers of the world's most widely used languages. The small grammar words are cmavo. Compounds, lujvo, are built by gluing root words together by fixed rules. Above all, the grammar is designed so that a computer can parse any correct sentence in exactly one way: no sentence is grammatically ambiguous.

Toki Pona: A Language With About a Hundred Words

The Opposite Experiment

Toki Pona ("good language") was created by the Canadian linguist and translator Sonja Lang and first appeared online in 2001, with a full description in her book Toki Pona: The Language of Good (commonly called pu). Where Lojban adds precision, Toki Pona removes almost everything: the original book describes around 120 words, and a later dictionary lists 137 "essential" words along with some less-used ones. The idea is inspired by simplicity and positive thinking: what happens to expression when you have to say complex things with very simple building blocks?

How It Works

There are no inflections, no articles, no plural and no tense. A word's job comes from its position, and a few small particles mark the structure. Meaning for anything not in the list is built by combining words, so a friend is a good person.

Does Language Shape Thought? The Sapir-Whorf Question

Loglan was meant to test what is called the Sapir-Whorf hypothesis or linguistic relativity: the idea that the language you speak influences the way you think. It comes in two forms:

For engelangs the question is practical. If the strong version were true, a language designed to remove ambiguity would make its speakers more logical, and a language with a handful of words would narrow their thinking. The evidence does not show either. What an engelang can fairly show is more modest: that a human language can work with a very different design, and that people can learn to use it. It tests what is possible in language, not how language shapes minds.

Hands-On Exercises