---
title: RFC: Development Roadmap
author: porpoiseless
---The semantics of a priori languages are often defined with complex, culturally specific terms. But these words come loaded with conceptual baggage, historical contingency, and oppositions both internal and external. They overlook the diversity of other ways of conceiving ourselves and the meanings of our speech, thought, and action.
- Words like "cognition" and "mental" are frequently used to define a word like "think", leading to circularity and obscurantism.
- Linguistic terms like like "Case" and "Aspect", may cover different semantic ranges across languages, and may be wholly unapplicable to others.
- Words for speech acts, which are highly dependent on the lifeways of a language community, may be baffling for outsiders who do not share that culture's system of values.
In short, our purposes require a great many words for describing how we think and what we mean when we say something. This document outlines a roadmap for making a language that is suitable for philosophical use outside of any given culture. That is not to say a culturally neutral language, but rather, a language built with a particular kind of culture in mind: namely a philosophical one.
Some time after now, for some time.
Draft.
- @porpoiseless [Cetacean needed]
Devil's in the details.
(TODO: cite Goddard's work on mental states and Wierzbicka's on enthnopragmatics as specific instances where NSM can help us conceive of lexemes in more fine-grained ways than the polysyllabic quasiclassical borrowings of English.)
We call a language "self-hosting" if it contains some minimal, well-defined subset that can completely describe the use and meaning of all the rest. In practical terms: a self-hosting language can be used to write a complete grammar and dictionary for itself, without resorting to circularity or the importation of complex foreign terms.
Languages with this property benefit in many different areas:
-
(pedagogy) We can teach such languages beginning with primitive, universally accessible concepts. Someone who has learned how speakers of this language talk about those concepts can in principle learn anything else about its meanings/uses in the language itself.
-
(documentation) Instead of using polysemic English words--and more often than not, obscure Latinate coinages--we can use native linguistic terms as early as possible. Our linguistics terminology is different. We need not worry if we use a word like "mood", or "case" differently from its commonly accepted use and meaning among lay people, its technical use in the scientific community, or its specialized use among some particular school or discipline of linguistics.
-
(repair) If a language can describe itself, speakers can use these facilities to correct miscommunications and make explicit the normative use and intended meaning of any linguistic structure.
Fortunately, Wierzbicka and Goddard have already developed and tested a semantic metalanguage of the sort we require. The Natural Semantic Metalanguage (NSM) consists of 65 primitives, each with a set of allowed syntactic frames. A self-hosting language must have ways of expressing these primitives and their allowed combinations. More complex terms can be built and defined on this basis, within the language rather than outside of it.
The following is a list of the semantic primes of NSM:
| Category | Primes |
|---|---|
| Substantives | I, YOU, SOMEONE, PEOPLE, SOMETHING/THING, BODY |
| Relational Substantives | KIND, PART |
| Determiners | THIS, THE SAME, OTHER |
| Quantifiers | ONE, TWO, SOME, ALL, MUCH/MANY, LITTLE/FEW |
| Evaluators | GOOD, BAD |
| Descriptors | BIG, SMALL |
| Mental predicates | THINK, KNOW, WANT, DON'T WANT, FEEL, SEE, HEAR |
| Speech | SAY, WORDS, TRUE |
| Actions, Events, Movement | DO, HAPPEN, MOVE |
| Existence, Possession | BE (SOMEWHERE), THERE IS, BE (SOMEONE/SOMETHING), (IS) MINE |
| Life and Death | LIVE, DIE |
| Time | WHEN/TIME, NOW, BEFORE, AFTER, A LONG TIME, A SHORT TIME, FOR SOME TIME, MOMENT |
| Space | WHERE/PLACE, HERE, ABOVE, BELOW, FAR, NEAR, SIDE, INSIDE, TOUCH (CONTACT) |
| Logical Concepts | NOT, MAYBE, CAN, BECAUSE, IF |
| Intensifier, Augmentor | VERY, MORE |
| Similarity | LIKE/AS/WAY |
Some of the primes will have morphological exponents, but ideally all will have
lexical exponents as well. Even if we can indicate TWO with the Dual Number,
we need to have a word to name Dual as a possible value of the category of
Number.
Many of the categories and values already tentatively present in the language lend themselves to ready expression in NSM. For example tables for NSM explications of the categories of Number, Composition, and Disposition are given below.
| Number | NSM |
|---|---|
| Non-Count | some of something |
| Singular | one something |
| Dual | two somethings |
| Plural | many somethings |
The NSM definitions for number end up being concise and easily understandable. There's really no need to say more!
| Composition | NSM |
|---|---|
| Homogeneous | things of the same kind |
| Heterogeneous | things not of the same kind |
Note that because we have selected a parsimonious and general definition of Composition, we need not make reference to "physical similarity", which would preclude this category from describing non-physical things (except, perhaps, metaphorically). By talking about "kinds" of things, we can use this category to distinguish a collection of texts (or books, music, etc) by genre, regardless of the physical appearance of these things (or even their physicality, in the case of digital media).
(The table for Disposition is more complex since two binary distinctions are involved. FreeTNIL terms are found inside the table, while NSM explications compose the table's axes. And yes, these terms are horribly self-indulgent.)
| Disposition | do one thing | do many things |
|---|---|---|
| in one way | Dirigent | Divergent |
| in many ways | Convergent | Frangent |
Note that unlike Ithkuil's category of Affiliation, Disposition never changes the underlying type of the set it modifies. While Coalescent Affiliation is useful, it's a typing nightmare because it's the only value in its category that can change the type of the set.
Connectedness poses something of a problem. I have reservations about using
FAR, NEAR, and TOUCH, as the basis for their definitions:
- Fused Connectedness implies more than touching: it implies that the boundaries between the items have been disrupted--while the act of touching requires boundaries that can be in contact.
FAR,NEAR, andTOUCHimpose a physical metaphor that might be ambiguous or inapplicable to some types. (They at least require that a distance function and boundary conditions be defined for set members). What does it mean if we say that two thoughts or feelings are "far from" or "near to" or "touching" each other?- Perhaps some kind of logical connectedness can be chosen for this category, rather than using a spatial metaphor. Maybe one that lets us give Fused a semantics of "parts of one thing", filling the gap left by Coalescent but not changing the type (since we're still referring to the parts, and not the whole).
Envelope proved difficult to analyze, and I wasn't satisfied with the explications. I think this category, too, needs to be reorganized. Perhaps this can be done with typing in mind: separating out the spatial from the logical from the causal from the spatial from the temporal.
This and the general task of explicating the categories of FreeTNIL are the next step in the roadmap.
The path to a self-hosted language is clear.
- Define our categories and their values in terms of NSM,
- Create lexical items for NSM primes, at least those that aren't encoded simply in the morphology,
- Define further terms in NSM, or in the subset of FreeTNIL that serves as exponents of NSM terms, or in some formalization isomorphic to NSM but with syntax highlighting,
- Create on this basis terms for phenomenological states, speech acts, etc, needed for philosophical thinking.
| Vote | Name |
|---|---|
| +1 | @porpoiseless |
Draft.
- @uakci