A MATHEMATICAL DEFINITION OF THE LINGUISTIC RECURSION

Jipsy KALONJI MUTEBA

[email protected]

Abstract
In the Minimalist Program, the idea of recursion in Language is connected to
the idea of ‘merge’. Recursion is the central idea of the Computational Theory, and due to
different computertional models, it may appear in different forms. The core computational
mechanisms of recursion are computability, definition by induction and mathematical
induction. It is the last, that is, mathematical induction that is studied in this paper. The purpose
of this paper is to make available to the linguistic community the standard mathematical
definition of recursion and apply it to discuss linguistic data. As result, we have realized that
there are “soft universals” in human language, which are related to cognitive constructs
necessary to implement mathematical reasoning. The Mathematical Model Theory and the
Principles and Parameters Theory are instumental for the analysis of data.
Resumé
Dans le Programme Minimaliste, l’idée de recursion est relative à celle
d’association. La recursion est l’idée principale de la Théory Computationnelle, et pour raison
de plusieurs modeles computationnels, elle apparait sous des formes diverses. Les mechanismes
computationels universels de la recursion sont la computationalité, la définition par induction
et l’induction mathématique. C’est le dernier, l’induction mathématique, qui nous concerne
dans cet article scientifique. Notre objectif est de fournir à la communauté linguistique une
définition mathematique standard de la recursion et son application dans la discussion des
données linguistiques. Comme resultat, nous avons trouvé qu’il ya des « elements universels »
dans le language humain, lesquels sont liés à des constructions cognitives necessaire dans la
mise en place d’un raisonnement mathématique. La Théorie Mathématique Modele and la
Théorie des Principes et Parametres sont instumentals dans l’analyse des données.

0. Introduction.

The concept of recursion is central to the study of Language as part of the brain.
But it is often badly used because the definitions povided by non-linguists are either wrong or
phrased ambiguously. Because recursion is a central notion in linguistics, it has to be made clear
 2

in the linguistic context. The notion of recursion is easy to define mathematically – and a
language is recursive if it is possible to implement the necessary mathematical concepts in it.

This paper explores similarities between some basic structures of mathematics

and linguistics. It is interested in the extent to which it is possible to encode the language of
mathematics in natural human languages. And here we use the word “mathematics” in a
cognitive structural sense that can be made precise using a branch of mathematics called
“Mathematical Model Theory”. Since the formalism of Mathematical Model Theory can
encode mathematical recursion, any natural language that can implement this theory, a-fortiori
implements recursion. And we in fact use this theory to define recursion. Mathematics shows
clearly that recursion is a faculty that helps language as an operating system (Universal
Grammar) to produce an indefinite number of possible sentences.

The point 1 introduces some standard mathematical formalisms and discusses what the
basic elements of mathematical statements are. This properly belongs to the field of
mathematical model theory. This naturally leads to point 2; here, we discuss “soft linguistic
universals” and give a more precise definition there. But, roughly speaking, by “soft
mathematical linguistic universals” we mean properties that are shared by most, but not
necessarily all, human languages. These are common structures, procedures, and rules. This
commonality is not necessarily genetically based but may arise from shared experience and
general cogitative traits of the human mind, or from social structures as well.
The point 3, turns around mathematical recursion. Its treatment is shows that the notion
of infinity is central to recursion. It also discusses discrete infinity. The point 4 applies
mathematical recursion to the linguistic setting. The point 5 summarizes conclusions. It is
worth noting that very little of this is new. But we think that it is important to bring together
the fields of mathematics and linguistics as both have much to offer to each other.

1. Mathematical Formalism
1.1. The Notion of Limit

I begin my discussion with an illustration from calculus. In what follows, we

shall let f(x) be a real valued function of a real variable x; defining this notion precisely is in
 3

itself instructive but we shall omit such a discussion in the interests of brevity. Newton would
have said:

(1)
To say that L is the limit of f as x tends to a means that if x is infinitesimally
close to (but different from) a, then f(x) is infinitesimally close to L. This
definition, when formalized, leads to a branch of mathematics called
nonstandard analysis. But as it stands, one may not verify if this definition is
true or false since it is not operationalizable.

Instead, we replace this definition by the following:

(2)
To say that L is the limit of f as x tends to a means that
gives any positive number ε, then there exists a positive
number δ so 0<|x-a|< δ implies |f(x)-L|< ε.

Any elementary treatise on analysis provides many examples using this

definition to prove mathematical assertions – this is an operationally useful concept. Rather
than belaboring the point, I refer, for example, to Ross (1980) for further details and examples.

1.1. Mathematical Model Theory

The English language plays a central role in the definition (2). However, one
can go yet one step further, and be even more very formal and remove this dependence and
render it symbolically. The following definition would be appropriate for a course in
nonstandard analysis (see Robinson (1996) or Kanovei et al (2004)) – the point being to
abstract the essential logical structure involved:

(3)
{L=lim(a,f)} ⇔{{∀ε∈R} ∧ {ε>0}}⇒ {{{∃δ∈R}∧{δ>0}} ⇒ {{{∀x∈R}∧{0<|x-

a|}∧{|x-a|<δ}} ⇒ {|f(x) -L|<ε}}}.

 4

It must be admitted, in gaining universality and precision, definition (3) has

lost ease of understanding and takes much effort to deconstruct. The study of mathematical
statements and arguments from this point of view belong to a field called Mathematical Model
Theory.
1.2. Basic Elements of Mathematical Statements
The basic elements in definition (3) and in common mathematical arguments
can be outlined in the following way:
i. Constants :
One can distinguish several subclasses – the following list is not exhaustive:
(a) Primitives: 0 means “zero” and “R” means “the real numbers”.
(b) Relations: “<” is a binary relation; a<b means “a is less b”.
(c) Functions: “|a|” is a 1-argument function whose input is a real number and
whose output is the absolute value of that real number; “lim” is a 2-
argument function whose first entry is a real number and whose second
entry is a real valued function; it is not defined for all a and f.
ii. Variables: x, a, L, ε, f, and δ.
iii. Quantifiers: ∀ means “for every” or “given any”; ∃ means “there exists”
or “one can find”.
iv. Implication: ⇒ means “implies that” or “such that”; ⇐ reverses the
implication and means “is implied by”; ⇔ means both implications, i.e. “if
and only if” or “equivalently”.
v. Negation ¬P means “not P”, i.e. “the statement P is false”.
vi. Conjunctions: ∨ means “and” while ∧ means “or”.

This is not a minimal list since, for example, the assertion {P⇒Q} is logically

equivalent to the assertion ¬{P∧¬Q} and {P∧Q} is logically equivalent to ¬{{¬P}∨{¬Q}}.

Thus, one can eliminate ⇒ and ∧ from the list of basic elements. And, of course, variables and
substitution are closely related but they are not identical concepts. What is important is that one
can in principle write down such formulas and admissible methods of argument very formally.
A set S of axioms (mathematical statements) is said to be “consistent” if there exists a model
satisfying these axioms. This is why the field is called “model theory”. It is a fundamental result
in the subject that only finite constraints pertain, i.e. an infinite set S of axioms is consistent if
and only if every finite subset T of S is consistent. As the models for each subset T may be
 5

different, it is necessary to “piece together” these different models using an ultra-filter to

construct a single coherent model for S. Note that there can be many models for a given set of
axioms; choosing a “non-standard” model for the real numbers permits one to discuss
infinitesimals and thereby make the definition (1) mathematically precise and rigorous.
2. The Linguistic Processing System
2.1. Language
Chomsky (1995: 14) suggests that language is an innate faculty, that is, humans
are born with a set of rules referred to as UG. The UG is the basis upon which all human
languages build. Chomsky makes it clear:

The language faculty has an initial state, genetically determined;

in the normal course of development it passes through a series of
states in early childhood, reaching a relatively stable steady state
that undergoes little subsequent change, apart from the lexicon.
To a good first approximation, the initial state appears to be
uniform for the species. Adapting traditional terms to a special
usage, we call the theory of the state attained its grammar and
the theory of the initial state Universal Grammar (UG).

Children learn a given language as a normal course of development as they are

facilitated by UG. When a child begins to listen to his/her parents, he/she will unconsciously
recognise which kind of a language he/she is dealing with – and he/she will set his/her grammar
to the correct one – this is known as ‘parameter setting’. He/she knows intuitively that there
are some words that behave like verbs, and others like nouns, and that there is a limited set of
possibilities as to their ordering within the phrase. This is not information that he is taught
directly by the adults that surround him/her, but information that is given. This set of language
learning tools, provided at birth, is referred to by Chomsky as the Language Acquisition Device
(LAD). Akmajian et al (2010) elaborate on how children use LAD to acquire phonological,
morphological and syntactic competence.

The evidence indicates that children, in fact, absorb a good number of sentences
and phrases and abstract rules from them and create their own grammar which they then apply
to create new utterances they have never heard before. Over the years from 2-7, when language
is mastered, children constantly adjust their grammar until it matches that of the adult speaker
population. This critical period between the ages of 2-7 suggests that first language learning,
like walking, is an innate capacity of human beings triggered by a level of development more
 6

than feedback from the environment. That is, so long as a child hears a language – any language
– when he/she reaches this critical period, he/she will learn it perfectly. Therefore, any child
not hearing language during this period would not learn to speak. This is known as critical
period hypothesis.

2.1.1. Essentialism :

Noam Chomsky has been called the intellectual ancestor of linguistic

essentialism, which aims to identify the intrinsic properties of language per se. Linguistic
essentialism is interested in postulating universals of human linguistic structures, unlearned but
tacitly known, that permit and assist children to acquire human languages. It prefers finding
surprising characteristics of languages that cannot be inferred from the data of usage, and are
not predictable from human cognition or the requirements of communication.

2.1.2. I-language and E-language:

Chomsky (1986) introduced two technical notions of language: E-language and I-

language. ‘E’ stands for ‘extensional’ and ‘external’ and E-language refers to language data or
corpus which is external to the mind. ‘I’ stands for ‘individual’, ‘internal’, and ‘intensional’,
and I-language means internalized language – the properties of the mind of individuals who
know them. (Chomsky, 1995: 15-16) As a generative essentialist sees it, I-language is a state
of mind/brain. Chomsky rejects Elanguage as undeserving of study and suggests I-language as
the only suitable object of study in linguistics. I-languages can and should be studied in
isolation from their external environments. Chomsky states:

For H to know L is for H to have a certain I-language. The

statements of the grammar are statements of the theory of
mind about the I-language, hence structures of the brain
formulated at a certain level of abstraction from mechanisms.
(Chomsky, 1986: 23)

He adds:

This branch of the study of language (generative grammar) is

indeed marked by an absence of any role for community and
culture ... There is nothing of any significance known, at least
to me, about community and culture that relates to these
questions about the nature of a certain biological system.
(Chomsky 2000: 42)
 7

This conclusion puts Chomsky in wrestling ring facing non-essentialists, who

think that language is a social-functional or external phenomenon. Michael Dummett has
thought language as a social custom, Esa Itkonen as social rules and David Lewis as social
convention. Piaget and Vygotsky also laid the primary emphasis on the role of social or
environmental factors in language development. Largely influenced by Wittgenstein, ordinary
language philosophers H. P. Grice, J. L. Austin and J. R. Searle postulated their pragmatic
theories like implicatures, conversational maxims and speech acts based on the practical use
of language (Levinson, 1983). In fact, ‘there are various views about the nature of language
and meaning that can be labelled externalist and Chomsky has been critical of them all’
(Bezuidenhout, 2006: 129). His conviction even pitted him against descriptive linguistics of
Leonard Bloomfield and structural linguistics of Ferdinand de Saussure. In his scientific
capacity, Chomsky views language as a biological organ or device. In this paper I look at
language as a biologic endowment.
2.1.3. Universal Grammar

According to Chomsky, Universal Grammar (UG) is the system of principles,

conditions, and rules that are elements or properties common to all languages – the essence of
human language. All human beings share part of their knowledge of language their common
possession regardless of which language they speak. The rules of UG provide the basic
blueprint that all languages follow. (Chomsky, 1976: 29; Cook & Newson, 1996: 1-2; Fromkin,
Rodman and Hyams, 2007: 25) UG theory attempts to clarify the relatively quick acquisition
of the mother tongue on the basis of minimum exposure to external input. Learning would be
impossible without universal language-specific knowledge. Chomsky argues that it is reasonable
to suppose that UG determines a set of core grammars and that what is actually represented in the mind
of an individual even under the idealisation to a homogeneous speech community would be a core
grammar with a periphery of marked elements and constructions. (Chomsky, 1982: 8)

He adds that human babies are born with the core linguistic sense common to
all language, which helps them to acquire any specific language from the environment.
According to Chomsky, the language faculty is part of our biological endowment, and as such
it is largely genetically determined. Neuropsychologist Eric Lenneberg in his Biological
Foundations of Language (1967) lends support to Chomsky’s view. He says the capacity to
learn a language is indeed innate, and, like many such inborn mechanisms, it is circumscribed
in time. If a child does not learn a language before the onset of puberty, the child will never
master language at all, as claimed in the critical period hypothesis.
 8

The proposal of universal grammar puts Chomsky in the core of rationalist

tradition. Chomsky’s proposal bears an affinity with the concept of archetype theorized by
noted Swiss psychiatrist Carl Jung. According to Jung, human beings are born with certain
inherited modes of functioning rooted in collective unconscious, referred to as archetypes.
Archetypes are conceived as innate neuropsychic centres possessing the capacity to initiate,
control, and mediate the common behavioural characteristics and typical experiences of all
human.
2.1.4. Competence and Performance

Competence enables people to generate all possible grammatical sentences.

Performance is the transformation of this competence into everyday speech. Chomsky
proposed that linguistic theory should explain the mental processes that underlie the use of
language. That is, the subject matter of linguistics will be competence, not performance. In the
same way, recursion a competence of Language as Part of the brain. It's use in E-language
depends on the performance of speakers. The failure of speakers of some languages to use
recursion is due to environmental disturbances and memory limitations. Despite this situation,
their brains still have the recursive faculty and they can make their languages recursive or learn
to speak recursive languages easily.

2.2. Linguistics
The very first question that daunts anyone confronted with the study of
linguistics is: What is linguistics? Linguistics is defined as a scientific study of Language. A
further question, consecutive to the first is: What is Language? This question has been
answered in a variety of ways throughout the history and subfields of the linguistic science.
For example, a system of communication, a medium for thought, a vehicle for literary
expressions, and so on (O’Grady, Dobrovolsky & Aronoff 1991: 1).

However, these definitions seem to say what language is used for rather than what
it actually is. On this ground, following Chomsky, we can view language (I- Language) as
part of the brain to which Chomsky refers as LAD (Language Acquisition Device). This
inborn linguistic competence is made up of a finite set of principles and allows any normal
human being to speak any natural language of the world (I-language). Besides, Chomsky
refers to the native speaker’s performance or E-language (Externalized language). For more
details consider 2.1.3 above.
 9

2.3. Mathematical and Linguistic Universals

The term “universals” in the linguistic field usually refers to traits that arise
solely from the genetic endowment. The notion of genetic determinism on language
acquisition is now part of folk-psychology, and it is widely popular in the fields of linguistics
and of language acquisition. As an effect, the expression “Language Universal” is implicitly
understood as “Language genetically-determined trait”. I can call this concept Universal
Grammar. In this paper, we choose to use the expression “soft linguistic universals”, implying
there are principles applied to all natural languages. (Bates 1979). Those constrictions,
together with some hard-wired genetic general cognitive factors, determine that languages
share some structural traits (e.g. the seven modules of Language). One might add to that, that
language acquisition is also a well-defined problem for all human babies, so that complex
human neural nets may come to similar and stable solutions for it.
In this context, generalized structural properties of natural languages wouldn´t
be a surprise, and the properties shared would be soft universals. Thus, I hope to bypass
entirely questions concerning Universal Grammar. Similarly, one could ask whether
Mathematical Model Theory is only a formal cultural development. This is more or less akin
to the philosophical questions:
i. Do mathematicians invent or discover mathematics?
ii. Is mathematics an innate reality or a human construct?
iii. Do the objects of mathematics have an intrinsic existence, which is extrinsic
to humanity?

There are constants, variables, quantifiers, implication, negation, and conjunctions

comprising the admissible statements of Mathematical Model Theory. Linguistically speaking
words and phrases such as “Chicago”, “Peter” “belong to”, “is the daughter of”, “windy”, or
“city” are all constants. Thus, the sentence “Chicago is a windy city and Emily is the daughter
of Peter” could be rendered as:

(4)
{{Chicago ∈ Windy}∧{Chicago ∈ City}∧{Emily∈Daughter (Peter)}}.
 10

Pronouns such as “he”, “it”, “she”, and “they” and Agr-elements in pro-drop
languages are clearly variables and admit substitution. Pronouns such as “some” and “all” are
not variables however; they are closely linked to quantifiers. On the other hand, some verbs
are constants like “belongs to” or “is a member of the set” (∈).
It is worth noting that the existence of a body of mathematical stuctures in a
natural language which uses logical reasoning and the axiomatic method is a sufficient
demonstration of the implementability of Mathematical Model Theory in that establishes that
particular Language is recursive.
2.3.1. Implementation
If there exist languages in which Mathematical Model Theory cannot be
implemented, this is an interesting feature that distinguishes such languages from other
languages with which I am familiar. And the phenomenon is not purely modern. There are
modern questions worth pursuing. Research by Dehaene et al (2008) provides a tantalizing
insight into Amazonian Indigene cultures:

“Our results suggest that all humans share the

intuition that numbers map onto space, but that culture specific
experiences alter the form of this mapping – and that a logarithmic scale
is characteristic of infants with a shift to a linear mapping occurring later
in Western children.”

Recursion is a faculty of Language but speakers of some natural languages fail

to apply it in their speeches. The fact that speakers of non-recursive languages can learn and
speak correctly languages in which recursion is present is a proof that recursion is innate (a
property of Language as part of the brain). Geometry constitutes a core set of intuitions present
in all humans, regardless of their language or schooling.
2.3.2. Efficiency
Concerning the efficiency of implementing mathematical constructs, Clark
(2006) argues that embodied agents use bodily actions and environmental interventions to
make the world a better place to think in. Where does language fit into this emerging picture
of the embodied, ecologically efficient agent? One useful way to approach this question is to
consider language itself as a cognition-enhancing animal-built structure. By materializing
thought in words, we create structures that are themselves proper objects of perception,
 11

manipulation, and (further) thought.” This suggests strongly that embodying basic logical
processes (such as exemplified in mathematical logic) in language enables “perception,
manipulation, and (further) thought”. Clark (2006) quotes Dehaene and coauthors as
presenting a compelling model of precise mathematical thought that reserves a special role for
internal representations of language-specific number of words.
3. Mathematical Recursion
3.1. The Peano axioms
Mathematical induction can be undersdood thanks to the following example
borrowed from Ross (1980).
(5)
“We denote the set {1, 2, 3...} of all natural numbers by N. Each natural
number n has a successor, namely n+1. One has the Peano axioms for this set:
N1) 1 belongs to the natural numbers.
N2) If n is a natural number, then its successor n+ 1 is a natural number.
N3) 1 is not the successor of any natural number. In particular, 0 is not a natural
number.
N4) If n and m are natural numbers which have the same successor, then n= m.
N5) A subset of the natural numbers which contains 1 and which contains n+1
whenever it contains n must be all of the natural numbers.”
Of course, this begs the question “what is a set” and that can be a crucial point
in non-standard analysis. One can also express the Peano axioms symbolically – and with
considerable loss of clarity – as follows:
(6) N1) {1 ∈ N}.

N2) {{n ∈ N} ⇒ {n+1 ∈ N}}.

N3) {{∃n ∈ N} ∧ {n+1=1}}.

N4) {{{n∈N}∧{m∈N}∧{n+1= m+1}}⇒ {n= m}}.

N5) {{{S⊂ N}∧{1∈S}∧{{n∈S}⇒{n+1∈S}}} ⇒{S= N}}.

The Peano axioms form the basis for mathematical induction and hence for
recursion. For each n ∈N, let P(n) be a list of statements or propositions that may or may not
be true. The principle of mathematical induction asserts that all these statements are true
provided P (1) is true and provided P(n+1) is true whenever P(n) is true.
 12

3.2. Recursive functions.

A function f can be defined recursively. For example, one sets n! =1·2·...· n
(read n factorial). The difficulty, of course, is in the “...”. One knows intuitively what is meant
by this expression. Clearly one just multiplies together the first n integers. But this is not a
mathematical definition. Rather, recursively, one uses the Peano axioms to define f (1): = 1
and then recursively (or inductively) sets f(n): = n · f(n-1).
There is an extensive literature concerning recursive functions. Such functions
(in principle) can be implemented on a computer. There exist non-recursively definable
functions – the literature is extensive and we give just a few references to provide a flavor.

3.3. Infinity

One can also give a precise definition to the meaning of ∞ (infinity) in this
regard – the notion of “discrete infinity” in linguistics is perhaps considerably less precise.
We will spare the reader more symbolic expressions and proceed using ordinary mathematical
notation. We say that a sequence s(n) of real numbers tends to ∞ if it increases without limit.
More precisely: A sequence s(n) of real numbers tends to ∞ if given any real number K, there
exists an integer N so that if n>N, then s(n) >K as explained in (6).
4. Linguistic Recursion : Mathematical Induction
4.1. Symbolism and Recursive Processing
The question of symbolism also is fundamental. Are humans first of all
symbolic, then subsequently linguistic, then recursive, and finally mathematical? Being first
symbolic (young children are not fully symbolic) has some implications. One could argue that
some of the basic principles of mathematical model theory are only symbolic. But recursion
enters since a symbol obtains all the constancies across the differences in the instances it refers
to. And of course, the notion of substitution into variables enters. And being symbolic, as
humans are, they can manipulate the symbols, as every other thing is manipulated This is, of
course, exactly how computer programs act by manipulating themselves and by making no
distinction between the data and the program. Thus, at some moment, inner symbol
manipulation becomes recursive, and at that moment, meta-symbols would be created.
4.2. Recursion in Language
Recursion plays a crucial role in most analysis of language. Marcus (2001)
notes any recursive scheme must have a set of primitives, a way of combining those primitives
 13

to form new complex entities, a way of ensuring that the arrangement of the elements matters
(for example so that 12 is not 21) or that ‘the cat is on the map’ is not ‘the map is on the cat’
and a way of allowing new complex entities to participate in the combinatorial process.”
Marcus (2001) also remarks in speaking of different models of human cognition that each of
these proposals turns out to implement the same machinery as the symbolmanipulation
account of recursion. Each of these models includes a systematic difference between atomic
and complex units, a way of combining these units to form new complex units, and a means
by which new complex units may in turn serve as input to further combinations.” One type
of recursion in language is mentioned in Radford (2004: 69), which I will term structural
recursion. Structural recursion is a view of recursion that deals with structures embedded in
other structures.

Recursion is the ability of language to encode or operationalize in an efficient

way certain universal modality. To greatly over simplify: we believe that languages have the
soft universal linguistic possibility to encode (i.e. implement) the fundamental logical
structures of Mathematical Model Theory. In other words, these basic mathematical logical
structures are reflected in language. Note that I am not speaking of numbers or geometry here
but rather the basic logical structure of mathematics (i.e. mathematical model theory as
discussed by Robinson [1996]).
The process of language learning itself is recursive in this sense, as is complex
language processing by adults. This relate to the production of discourse or to writing, for
example. But Music processing can also be recursive, and social cognition, and visual object
deconstruction, and perhaps much of conscious processing engaged in complex problem
solving. It is possible to define the notion of “mathematical truth” for formulas in
mathematical model theory recursively. Human implicit recursive processing can perhaps be
expressed as something like: “do it again”, “take now those new elements you obtained
(however abstract and symbolic they already are) and apply to them the same logic you had
used before, and see what comes out now”. That is, of course, very similar to the mathematical
notion of a mathematical recursive function.
4.3. Symbolic Communication, Numbers, and Discrete Infinity
Note that we are not asserting that language is just an implementation of
mathematical model theory – far from it. The elements that we identify as corresponding to
mathematical model theory are surely simply the basics; they are necessary underpinnings of
 14

but are far from the sufficient conditions for a successful human language. Anyone with even
a mild ear for language can distinguish spoken Korean, Japanese, and Chinese without
understanding a word of these languages simply from the sound of the language; this facility
to distinguish between linguistic groups has also been demonstrated vary early in children.
So, it is clear that language has a structure, a rhythm and a poetry far removed from the logical
underpinnings that we have identified here. But with that essential caveat, we believe that a
careful discussion of some soft universal elements of language from a mathematic model
theoretic viewpoint is likely to be a fruitful one. And we hope this approach will be a felicitous
undertaking, which will avoid the sterile formalism of previous attempts to apply symbolic
logic to linguistics. Tomasello (2003) notes:

“First, and most importantly, human linguistic

communication is symbolic … human symbols are aimed
at the attentional and mental states of others ... human
beings use their linguistic symbols together in patterned
ways, and those patterns, known as linguistic
constructions, take on meanings of their own – deriving
partly from the meanings of the individual symbols, but,
over time, at least partly from the pattern itself.”
Also relevant is Pinker and Jackendoff (2005) that states that more striking is
the possibility that numbers themselves (beyond those that can be subitized) are parasitic on
language – that they depend on learning the sequence of number words, the syntax of number
phrases, or both”. It also notes that recursion consists of embedding a constituent in a
constituent of the same type for example a relative clause inside a relative clause which
automatically confirms the ability to do adlibitium.
Hauser, Chomsky, and Fitch (2002) state: “The computational mechanism of
recursion is recently evolved and unique to our species ... only those mechanisms underlying
FLN – particularly its capacity for discrete infinity – are uniquely human.”They also state:
“Recursion is agreed by most modern linguistics to be an indispensable cores computational
ability underlying syntax and thus language ... there are no unambiguous demonstrations of
recursion in other human cognitive domains with the only clear examples (mathematical
formulas, computer programming) being clearly dependent on language.”
4.4. Types of Recursion in Linguistics
 15

All the types of recursion are squeezed in two main types viz structural
recursion and functional recursion. The definition provided by PJ (2004: 203), makes the
distinction between the two definitions of recursion explicit: "Recursion refers to a procedure
that calls itself, or to a constituent that contains a constituent of the same kind”. The former
type of recursion is functional recursion, and the latter is structural recursion.
The two definitions provided by Parker (2006:3) neatly detail both functional
and structural recursion. Structural recursion occurs when an object is embedded within another
object. Whether it is embedded at the centre or edge is a secondary concern – what matters for
the definition is that the object is contained by another object of the same type. Functional
recursion follows the same principle in that a process is embedded within another process. Just
like with objects, processes can be embedded within other processes. A process embedded "in
the centre" of another process indicates that the first process starts, and is interrupted midway
to allow the embedded child process to start. Only when all the embedded child processes have
completed, is the first process allowed to continue.
4.4.1. Linear Recursion
A linear recursive structure is a structure that only makes a single call to itself
each time the function runs (as opposed to one that would call itself multiple times during its
execution) as in (7) and (8)
(7) 1+O=1
There is recursion since 1 is manipulated to get one. The factorial function is a
good example of linear recursion in Mathematics. Let us apply it to the following Phrase
Structure.
(8) a. X’’→ Spec X’
b. X’→X°

X’ in (7a) is manipulated to get X” and X° in (7b) is manipulated to get X’.

Therefore, there is linear recursion from the mathematical perspective. The value of merge at
XP step is defined by the value at step XP-1 or X.
4.4.2. Tail Recursion
It is a form of linear recursion. The recursive call is the last thing the function
does. Often, the value of the recursive call is returned. As such, tail recursive structures can
often be easily implemented in an iterative manner.
 16

(9) a. Eat bananas (VP)

b. VP→V, NP (The value of VP is defined by V, NP)

c. V”

Spec V’

V° N”

Eat N’

N°

bananas

V’ is the tail of V”; That is, V” takes its interpretation or definition from
V’. V° is the tail of V’.
4.4.3. Binary Recursion
Some recursive categories do not just have one call to themselves, they
have two or more. Categories with two recursive calls are referred to as binary
recursive structures.
(10) a. Jipsy knows that Bob knows that Chomsky is a linguist.
b. [[Jipsy knows] IP1 that [Bob knows] IP2 that [Chomsky is a linguist] IP3] IP
IP1 calls the same category twice (IP2 and IP3). Thus, an instance of binary
recursion.
4.4.4. Exponential Recursion
An exponential recursive category is one that, if you were to draw out a
representation of all the category calls, would have an exponential number of calls in relation
to the size of the data set. It is like binary recursion but here a category has multiple calls to
itself.
 17

5. CONCLUSION
It follows from Mathematical Model Theory that recursion is part of the species
endowment (U.G), and thus universal. But not using the tools provided by Language (as part
of the brain), a language might well not be able to exercise this facility(recursion); thus,
existence of non-recursive languages. This paper shows clearly that the “Soft Mathematical
Universals” of Language, which have been observed previously, are necessary to implement
the logical constructs of Mathematical Model Theory. The elements such as negation,
variables, constants, conjunction, quantifiers, and implication are soft universals of natural
languages precisely because implementing recursion is not possible if these elements are
absent. Both Linguistics and Mathematics are constructs of the same cognitive system; with
Language, the human cognitive system is recursive, and some soft universals of Language are
necessary prerequisites to ensure recursivity in natural languages. From mathematical
induction, the different types of linguistic recursion are: linear, tail, binary and exponential.
As in mathematics, in linguistics, the value of merge at XP step is defined by the value at step
XP-1 or X.

BIBLIOGRAPHY

1. Chomsky, Noam, 1957/2002, Syntactic Structures, Berlin: Mouton de Gruyter.

2. Chomsky, Noam, 1965, Aspects of the Theory of Syntax, Cambridge, Massachusetts: MIT Press
3. Chomsky, Noam, 1966, Cartesian Linguistics, New York: Harper and Row.
4. Chomsky, Noam, 1969, “Linguistics and Philosophy”, in Language and Philosophy: A Symposium,
Sidney Hook, (ed.), New York: New York University Press, 51–94.
5. Chomsky, Noam, 1976, Reflections on Language, London: Temple Smith.
6. Chomsky, Noam, 1977, Language and Responsibility, New York: Pantheon Books.
7. Clark, (1998), “Magic Words: How Language Augments Human Computation”,
Interdisciplinary Themes, Cambridge University Press 162–183.
8. Clark A., (2006), “Language Embodiment, and the Cognitive Nitch”, Trends in Cognitive
Sciences 10 370-374.
9. --------(2006), “Material Symbols”, Philosophical Psychology 19 291–307.
10. Davidson D., (2001), Subjective, Intersubjective, Objective, Clarendon Press
11. Davis M., (1977), Applied Non-standard Analysis, Wiley
12. Dehaene S., Izard V., Spelke E., and Pica P., (2006), “Core Knowledge of Geometry in an
Amazonine Indigene Group, Science 311 381–384.
 18

13. -------- (2008), “Log or Linear? Distinct

14. Brody F. and Vamos T., (1995), “The Neumann Compendium”, World Scientific Series in
20th Century Mathematics, 1. World Scientific Publishing Co., Inc. Intuitions of the Number
Scale in Western and Amazonian Indigene Cultures”, Science 320 1217–1220.
15. Everett D., (2005), “Cultural Constraints on Grammar and Cognition in Piraha”, Current
Anthropology 46 621–646.
16. Goddard C. and Wierzbicka A., (2002). Meaning and Universal Grammar: Theory and
Empirical Findings, vol 2, John Benjamins,
17. Hauser M., Chomsky N., and Fitch W., (2002), “The Faculty of Language, what is it, who has
it, and How did it evolve?”, Science 298 1569–1577.
18. -------- (2008), The Evolution of the Language Fcility: Clarifications and Implications, Press
Cognition.
19. .Hodgkin L., (2005), A History of Mathematics from Mesopotamia to Modernity, Oxford
University Press,
20. Kanovei V. and Reeken M., (2004), “Nonstandard Analysis, Axiomatically”, Springer
Monographs in Mathematics, Springer-Verlag,
21. Kazadi-Mukenge J., (2009), A Characterization of Ciluba, Mbujimayi: Presses Universitaires du
Kasai.
22. Marcus G., (2001). “The Algebraic Mind, Integrating Connectionism and Cognitive Science”,
A Bradford Book, MIT Press
23. Neugebauer O., (1951), The Exact Sciences in Antiquity, Oxford University Press
24. Peter N., (1951), Rekursive Funktionen, Akademischer Verlag
25. Pinker S. and Jackendoff R., (2005), “The Faculty of Language, What’s Special About It?”,
Cognition 95 201–236.
26. Premack D.G., and Woodruff G., (1978), “Does the Chimpanzee Have a Theory of Mind?”,
Behavioral and Brain Sciences 515–526.
27. Robinson A., (1996), Non-standard Analysis, Princeton University Press
28. Rose H.E., (1961), “On the Consistency and Undecidability of Recursive Arithmetic”, Z. Math:
Logik Grundlagen Math. 7 124–135.
29. Ross K., (1980). Elementary Analysis: The Theory of Calculus, Springer-Verlag
30. Stephan F. and Zeugmann T., (2002), “Learning Classes of Approximations to Non-recursive
Functions”, Algorithmic Learning Theory, Tokyo, Theoret. Comput. Sci. 288 309– 341.
 19

31. Stone V.E, and Gerrans P., (2006), “What’s Domain-Specific About Theory of Mind”, Social
Neuroscience 1 309–319.
32. Tomasello M., (2003), Constructing a Language, a Usage-based Theory of Language
Acquisition, Harvard University Press.