kandinsky_Thirty_(Trente)__1937

IEML, the Information Economy MetaLanguage, is a free and open-source artificial language with computable semantics. It is the result of 20 years of research, mainly at the Canada Research Chair in Collective Intelligence at the University of Ottawa.

The main purposes of IEML are (1) to provide a semantic metadata system on the Internet and (2) to build a bridge among classifications, ontologies, languages, and jargons that fragment digital memory.

FOR WHOM IS IT INTENDED?
IEML is a multidisciplinary project at the confluence of AI, data science, linguistics, digital humanities, and philosophy. Because the metalanguage IEML has computable semantics it will be of interest to people working in the fields of artificial intelligence, business intelligence, and data science. This metalanguage proposes new uses and theory of metadata also relevant to researchers in the fields of heritage conservation (libraries, museums), digital humanities, and data journalism. Finally, since IEML increases collective intelligence, it will be of interest to practitioners in knowledge management, collaborative learning, and digital communications.

In this day and age, semantic interoperability among databases, languages, disciplines, etc. is a problem for a lot of professionals and researchers in the above-mentioned fields. In addition, after several years of deep learning frenzy, there is a renewed interest in symbolic AI (or at least in a synthesis between statistic and symbolic AI), and IEML is a powerful symbolic tool.

UNIQUE FEATURES
IEML has the same expressive power as a natural language; but it is also a regular language whose syntax and semantics can be computed. In other words, IEML algorithms automatically determine semantic relationships between any two grammatical units of the metalanguage (i.e. between morphemes, words, simple sentences, complex sentences, texts). And IEML is translatable into any natural language.

As of today, there is no existing language which simultaneously has the following properties:
1. IEML is a philological language (according to Louis Helmsley’s definition), meaning that it has the same power of expression and translation as any natural language.
2. It is a regular language in the Chomskyan sense, its syntax being computable.
3. Its semantics can be computed, from the tiniest element being the morpheme, to the largest level being the text.

These three properties combined make IEML a powerful tool to ensure semantic interoperability, organize information, and model knowledge in a  flexible way in the age of the Internet of things, big data, artificial intelligence, and digital humanities.

From a mathematical point of view, IEML is a “topos,” i. e. an algebra (a regular language) in relation of morphism with a topology of semantic relationships.

COMPETITION
As of today, IEML has no equivalent. However, let’s discuss two somewhat similar projects – one technical (the Semantic Web) and one linguistic (Lojban).

IEML and the Semantic Web
In terms of long-term vision, IEML’s main competitor is the WWW Consortium’s “Semantic Web,” including the RDF and OWL standards. The Semantic Web project was formulated at the end of the 20th century and is based on the availability of inference engines and ontologies (rule-based systems representing domain knowledge) that were developed at the time as the “expert systems” in the 1970s and 1980s. However, at the time not all computers were interconnected and the problem of semantic interoperability was not as acute as it is today. Yes, the Semantic Web makes it possible to compute semantic relationships and make logical inferences, but only within an ontology. There are thousands of different existing ontologies and translations between ontologies must be done “by hand.”

The Semantic Web and IEML have some features in common. In particular, both aim to “represent knowledge” for automatic reasoning. But they are two different endeavors: the Semantic Web is a set of standards for data import/export XML (Extensible Markup Language) and the description of logical relations between data RDF (Resource Description Framework), OWL (Ontology Web Language). None of these standards have – like IEML does – the properties of a philological language. The Semantic Web standards have been designed for computing the truth of propositions in whatever natural language they have been expressed in, while IEML computes the semantic relations between any IEML morphemes, words, sentences and texts (and IEML may be automatically translated in natural languages). IEML and OWL operate on very different levels but are not mutually exclusive: you can write IEML texts in XML, RDF, or in OWL files.

During its two decades in existence, the Semantic Web has achieved a lot of useful applications. But it has also shown that problems in semantic computing and interoperability can’t be solved by logic and classifications alone. IEML, being both a formal and a philological language, computes, generates, and recognizes automatically an infinity of concepts and their semantic relations. In short, IEML is not a universal ontology, it is a language with computable semantics that can express any ontology.

The World Wide Web Consortium project represents semantic relationships by means of logical formalism and its nodes are arbitrary sequences of characters (like URLs). In contrast, IEML represents semantic relationships through linguistic formalism with nodes (USLs, uniform semantic locators) that are expressions of a single regular language. Inherent to IEML expressions, semantic relationships are represented here in a much more parsimonious way.

Finally, for the Semantic Web, URLs are the last layer of (physical) addressing of data and a conceptual addressing system is impossible or even reprehensible. In contrast, IEML proposes in the long term a universal addressing system for concepts on the Internet: USLs (uniform semantic locators).

IEML and Lojban
People who are interested in artificial languages always mention Lobjan as a possible competitor to IEML because it has a logical and regular grammar.

Lojban’s grammar is inspired by pre-existing logical formalisms and not by an original, flexible, effective, and general abstract algebra like IEML.
The meaning of Lojban words is solely defined by a correspondence with natural languages and not by a coincidence of signifier (syntax) and signified (semantic) paradigms as found in IEML. Moreover, in Lojban the level of morphemes does not meet the standards of a regular language. In contrast, IEML semantics are entirely generative. In short, IEML is the only artificial language that – once given its semantic primitives – defines its semantics by means of its syntax.
Finally, Lojban is made to be spoken while IEML is designed from the beginning as a scientific writing to serve as a semantic metadata system capable of indexing all aspects of human activity, that can be manipulated and “understood” by computers and to solve the problem of semantic interoperability.

PHILOSOPHICAL AND ANTHROPOLOGICAL PERSPECTIVE
The human species can be defined by its special ability to manipulate symbols. Each great augmentation in this ability has brought enormous economic, social, political, religious, epistemological, educational (and so on) changes.

There have been only four of these big changes. The first one is related to the invention of writing, when symbols became permanent and reified. The second one corresponds to the invention of the alphabet, Indian numerals and other small groups of symbols able to represent “almost everything” by their combination. The third one is the invention of the printing press and the subsequent invention of electronic mass media. In this case, the symbols were reproduced and transmitted by industrial machines. We are currently at the beginning of a fourth big anthropological change because the symbols can now be transformed by massively distributed automata in the digital realm. We still do not have invented the symbolic systems and cultural institutions fitting the new algorithmic medium. So my research in the past 20 years has been devoted to the invention of a symbolic system able to exploit the computational power, the capacity of memory and the ubiquity of the Internet.

This is the main motivation behind my work on IEML: I took up the challenge of inventing a symbolic system that makes the most of the new digital environment to serve human cognitive augmentation better.

Follow me and IEML on Twitter
For scientific publications, reports and other documents, look here.

Prof. Pierre Lévy, PhD., University of Montreal
Fellow of the Royal Society of Canada