A person's language competence is sometimes reduced to his mastery of a language system. In the following, this will be mentioned, too; but language competence comprises much more (Lehmann 2007). Its fundamental dimensions develop along the following distinctions:

Levels of specificity

The first distinction to be made inside language competence regards two levels of generality:

• Semiotic competence is the capacity of semiosis, i.e. to learn and use a sign system for cognitive and communicative purposes. This regards not only natural human languages, but also nonverbal and paralinguistic communication. This competence is, in the first place, based on very general human faculties:
  • man's anatomic and physiological equipment, which enables him to articulate and decode language sounds, to process utterances in memory etc.
  • cognitive and communicative abilities which are ultimately based on intelligence and the ability to entertain social relations.
  This is complemented by specifics distinguishing natural human languages from other systems of communication, above all by double articulation as well as the division of the significative system into the subsystems of lexicon and grammar.
  Semiotic competence is for the most part innate to a person. Just as for all other human properties transported in the genes, this implies that it is, on the one hand, a genetically anchored property of the species Homo sapiens, but on the other hand, individuals can differ markedly in its shape. As is well known, a person may have such a low communicative aptitude that pedagogical efforts in this regard are essentially futile.
• Language-specific competence is the mastery of the system of a particular language. The language system, in turn, comprises the significative and distinctive subsystems with their various components. It is, consequently, internally heterogeneous and therefore complicates the language competence by additional dimensions.
  Language-specific competence presupposes semiotic competence and is acquired entirely through learning.

Independent evidence for this level distinction is provided by gesture languages (or sign languages): Persons with congenital deafness cannot easily learn a vocal language and acquire a language-specific competence of the kind acquired by healthy people. They possess, however, the same semiotic competence as other persons; and this enables them in principle – i.e., apart from their deafness – to acquire a language-specific competence. This will then be the competence in a gesture language.

An issue that has attracted considerable theoretical dispute is the following: Semiotic competence is the prerequisite for learning and using any human language. Thus, on the one hand, it cannot include specifics that are limited to certain languages; but on the other hand, it must be sufficiently developed so that the individual language system can seamlessly connect to it and that language learning can directly build on it. After all, the latter is something that distinguishes humans from animals and is even easy for humans, at least in childhood. This raises the question of whether language universals are encoded in human genes. This is an empirical question that will ultimately be answered by human genetics. Until the answer is found, the following distinction is plausible:

• There are language universals which necessarily derive from some other circumstances - logical, physical, biological, functional, etc. This applies, for example, to the universals that govern word order patterns; these can be explained functionally. For these it is not necessary to assume that they are innate (beyond the biological endowment mentioned before), because they would apply anyway, even if they are not innate.
• In addition, all human languages may share properties that do not follow necessarily in this way, but are, as it were, arbitrary constraints on the human language faculty if compared with conceivable language faculties. Such constraints would originate in the specific evolution undergone by Homo sapiens and would therefore be coded in the genes.

The latter position has been postulated and developed into a theory of ‘universal grammar’ (Chomsky 1995). It has been known as nativism. However, there are neither candidates for innate universals which would have kept this status for more than twenty years of the history of science, nor are any methods known by which they could be identified.

Cognitive levels

The second distinction within language competence to be made here concerns two cognitive levels:

• Procedural competence is a bundle of capacities and skills in a specific domain. Procedural language competence is consequently the ability to think and communicate in a language. It may be subdivided into a set of skills derivable from a cross-classification of the two binary distinctions between productive vs. receptive direction, on the one hand, and oral vs. written medium, on the other (s. below). An important aspect of this competence is the fluency and correctness with which these skills are practised. This is the ability which is commonly examined in proficiency tests.
• Reflective competence is the capacity to control procedural competence in a certain domain. It presupposes reflection and knowledge of this procedural competence. Reflective language competence is consequently the ability to reflect upon language activity and, e.g., to explain to somebody else how it works. Reflective competence is recursive; a higher level of reflective language competence is linguistic competence. This is the ability which is commonly examined in a grammar test.

The two levels of competence are mutually independent to a considerable extent. On the one hand, there are persons who can make themselves understood in a language, but are in no position to utter a sensible sentence about it (for instance, in order to teach somebody the language). On the other hand, there are persons – typically, linguists – who possess a high level of reflective competence on a language (e.g., Latin) without being able to communicate in it.

For purposes of everyday life, procedural competence is basic. However, to perfectly dominate something in a human sense implies being able to assume responsibility for it; and this presupposes reflective control.

Two types of memory are responsible for the two cognitive levels of competence:

• Procedural memory stores routines of actions and processes that we master or that run automatically in us. The ability to ride a bicycle, e.g., is stored in our procedural memory. Its content is not accessible to introspection.
  Procedural memory is located in the brain area which is the oldest in evolutionary terms, viz. the brainstem and the cerebellum.
• Declarative memory stores propositional knowledge. For instance, the proposition that in riding a bicycle, one drives tiny wiggly lines to keep one's balance is stored in declarative memory. The content items are accessible to introspection, they can be rendered in language and shared with somebody else.
  Declarative memory is located in a younger brain area, viz. in the cortex and the diencephalon.

The two types of memory are of different evolutionary age; declarative memory is an achievement of Homo sapiens. This is why they are located in different regions of the brain.

Active and passive language competence

The last division to be discussed here concerns the mode or direction in which we activate our procedural competence, viz. the contrast between speech perception (comprehension) and speech production (speaking).

• Passive language competence or receptive competence is the ability to understand utterances.
• Active language competence or productive competence is the ability to produce and convey utterances.

Add to this mediative competence, i.e. the ability to convert messages from one language into another. It is a complex combination of the two directions of language competence, viz. the combination of passive competence in one language with active competence in another language.

Again, this division may be cross-classified with the different communication media. This results in the following modes of linguistic communication:

These modes generate a subdivision of language competence into modal skills that figure prominently in many language proficiency tests.

The relationship between active and passive language competence is asymmetric in several respects:

• For any language, passive competence properly and completely includes its active competence. This is true both for one's native language and for non-first languages. And it applies to all domains of the language system: the passive vocabulary is larger than the active one; we discriminate – particularly in a non-first language – more language sounds auditively than we distinguish articulatorily. We understand different dialects of our language which, however, we do not speak. Many people can read more or less, but cannot write.
• In language acquisition, passive language competence precedes active competence. This is necessarily so because the elements and procedures get into our head only by perception. In first language acquisition, the phase displacement between listening and speaking amounts to one year on average; and if the prenatal period is added, it is even longer.
• In phylogeny, too, the sensory and analytic capacities underlying the semiotic faculty are developed earlier than the pertinent motor and synthetic capacities. For instance, all higher animals can interpret indexical signs, e.g. the shadow of a raptor on the ground. By contrast, the ability to intentionally produce such signs – and a fortiori, iconic and symbolic signs – is limited and to a substantial extent emerges only in primates.

The asymmetry between active and passive language competence plays an important role in non-native languages. For many purposes, it suffices to dominate a second language passively. One of these is academic study or, more generally, collection of information. A particularly clear case is represented by ancient languages such as Latin or Sumerian, which many philologists and historians can read, while they lack all other modes of communication. Another such purpose is international correspondence: in academic spheres, “Erasmic communication” (after Erasmus of Rotterdam [1467-1536]) is widely used, where each correspondent uses his native language in production.