In order to speculate sensibly about what an exolanguage might be like, it is necessary to take a closer look at language on Earth in a number of separate but related steps.

Throughout this part of the book, references will be made to how some features of human language might be manifested in a possible exolanguage. In this way it is hoped that the speculation about exolanguages will be seen as grounded in established knowledge about language(s) on Earth. The question is not only how did language evolve on Earth but also how humans arose with the language faculty in the first place, so as to then possibly extrapolate from our evolution on Earth to that which might be found on exoplanets. In this context it will be worthwhile considering communication systems among non-human animals to relativise statements about human language. The goal is to establish structural and functional principles which hold for terrestrial communication systems and then consider how these might be realised on an exoplanet with intelligent beings.

20.1 What Is Language?

To begin with, one should distinguish carefully between languages, found throughout the world, and the human language faculty.Footnote ¹ A singular use without an article – ‘language’ – refers to the unique faculty humans possess, as part of their biological endowment, this enabling us to acquire a specific language, to understand and to use it (speak and/or sign it). Existing languages share a core of structure common to all, as dictated by the language faculty of humans (see Chapter 23 below on universal grammar).

Particular languages – English, Russian, Swahili, Hindi, Japanese, etc. – are products of the language faculty, when it has been active in the members of a social community over many centuries. The English language (in its many varieties), as we hear others speak it and as we see it written, is an external manifestation of speakers’ internalised language, which all native speakers possess in their brains and which they use for organising and manipulating their thoughts. In the technical parlance of the linguist Noam Chomsky, the ‘language of thought’ is an i-language (‘i’ stands for ‘internal’) and the language manifest in speech and/or signs is an e-language (‘e’ stands for ‘external’).

There is an order to the sequence of language faculty and languages. The language faculty is what allows actual languages to arise and develop. Languages are the products of the language faculty operating. Without this faculty there would be no languages. So, when considering humans from an evolutionary point of view, the first question concerning language is: ‘How did the faculty for language arise?’ Scholars are divided on this and there are two main camps, those who adhere to the continuity view, put simply: the language faculty arose slowly from a simple beginning and gradually grew more complex; and those who hold the discontinuity view, again put simply: the ability to speak rests on a sudden mutation in one single individual, which led to that person acquiring the ability to have language as we know it, with an internal hierarchical and non-linear structure (see Section 32.7 below for further discussion). The latter is a minority view, but one which is held by the most influential linguist of our time, Noam Chomsky, which in itself means that this view is considered seriously in all works on the origin of language.

20.2 The Purpose of Language

This might seem to be a straightforward matter: language is used to convey information or to express emotions; language is a system of communication used by humans. After all, we have the urge to communicate with others, expressed neatly in the German word Mitteilungsbedürfnis ‘wish or need to communicate’ (a term commonly used by the American linguist W. Tecumseh Fitch). This is true, but by no means the whole story. Communication is an external function of language, manifest in exchanges between humans. But language is also a vehicle for thought, an ability which we carry in our brains to organise and direct our cognition. In this respect the organisation of language is probably not primarily linear with a temporal ordering of elements. (I will return to this issue later, see Section 32.7.)

But there is a strong social component to language as well. It is used to maintain social relationships and to identify with a certain section of society. This means that all human languages have two sides: an internal structure, concerned with the organisation of linguistic information necessary for thought; and an external aspect for communication, where the manner in which language is expressed carries social significance. When one considers the first aspect, the internal organisation of language, one can see that in the course of human evolution our ability to speak would appear to have become self-contained and independent of our environment. Not only that, the levels within language, those of sounds, words and sentences, would also seem to have become largely independent but with connections linking them. This modularisation is a distinct advantage to the organisation and maintenance of language and is the reason for treating the levels separately in books on language.

20.3 Definitions of Language

Normally there is tacit agreement among linguists as to what constitutes language. However, when pressed on the matter, they find it difficult to come up with a single definition which satisfies everyone. By their very nature, definitions of language try to compress into a single sentence the essential elements that characterise it.

The American anthropologist and linguist Edward Sapir (1884–1939) defined language as ‘a purely human and non-instinctive method of communicating ideas, emotions and desires by means of voluntarily produced symbols’ (Reference SapirSapir 2004 [1921]: 3). This definition highlights the fact that language is used for communication by the use of symbols, usually auditory, but also possibly visual, as in sign language. Whether language is non-instinctive, as Sapir maintains, is a contested issue.

There are many other definitions of language and that by Noam Chomsky (1928–) stands out by emphasising the centrality of grammar and its ability to produce a potentially limitless number of sentences: ‘I will consider a language to be a set (finite or infinite) of sentences, each finite in length and constructed out of a finite set of elements’. The common ground shared by the very different definitions just given can be usefully summarised.

20.4 Design Features of Language

Definitions of language are by their very nature single-sentence renderings of what scholars see as the essence of the subject matter. But some scholars have gone beyond that to list and discuss features which they see as true of all human languages. Such features would have arisen before the last dispersal of Homo sapiens from the continent of Africa about 70,000 years ago, as they are found to hold everywhere, irrespective of the observable differences between the world’s languages in pronunciation, grammar and vocabulary.

A well-known classification of the design features of language was offered by the American linguist Charles Hockett (1916–2000), see Reference HockettHockett (1960), and represents an effort to comprehensively list the essential features of human language. The following list is based on the original one by Hockett but has been expanded to include further relevant information.

General Features

Relationship of words to concepts/objects is arbitrary: ‘Arbitrary’ in linguistics denotes a relationship between linguistic signs (words) and what they stand for (concepts, which typically refer to objects in the outside world). This relationship is not fixed or determined by the nature of the objects. In any given language, the relationship between word and concept/object is set by convention, for instance there is no reason why a female bovine animal should be referred to as cow [kau] in English, bó [boː] in Irish, vache [vaʃ] in French or korova [kʌˈrovə] in Russian. Nonetheless, to speakers of these languages, their word seems to be entirely appropriate for this animal.

Stimulus-free: As opposed to most animal communication systems, human language does not need a trigger such as danger or the search for food or the desire for procreation. In essence, we can speak without any external motivation.

Structure-dependent: Language does not consist of a string of random elements. The sounds and words of language are arranged in a certain meaningful order determined by the rules of the language.

Duality of patterning: A major organisational principle of language is that it involves two levels of structure, one of units and one of elements used to build these units (Figure 20.1). Words consist of sounds which in themselves have no meaning. For instance, one cannot say that the sounds /p/ p, /ɛ/ e or /n/ n have a meaning on their own but the combination /pɛn/ pen does. The same applies on a higher level: words combine to form sentences which they do not form on their own.

Figure 20.1 Duality of patterning

Discreteness: This requires that one has an exact realisation for each sound in a language. It represents the essential difference between noise and the sounds of human languages. Noise can vary at random but the sounds of a language must be kept apart clearly, that is, they are ‘discrete’Footnote ² in the technical sense. In English, one cannot use a sound which is intermediary between /p/ and /b/ as this would not be sufficiently separate from both of these. This applies equally to vowels. Again, in English one must distinguish clearly between vowels as, for example, in bid, bed, bad, bud, booed, bard, bide, bowed.

Productivity: The number of utterances one can make in a language is not limited: new sentences are produced by taking one of a limited set of sentence structures and filling it with words from one’s vocabulary. By these means, one can produce a theoretically unlimited set of sentences (see the discussion of digital infinity in Section 32.7 below).

In word formation, one can also see this principle at work. Take the example of the ending -wise, which is used to make adjectives from nouns. This can be applied productively to virtually any noun, irrespective of whether the new word already exists or not, such as Computerwise the department is well equipped. What enables productivity is the application of rules to any input element to gain a new structure, be it from grammar or vocabulary.

Gaps and prohibitions: A gap is a permissible form in a given language, which happens not to be attested. Prohibition refers to the exclusion on principle of some forms from a language. Consider the form blick, which is a possible (but unattested) word in English and bnick, which is not permissible because no English word can begin with a stop and a nasal.

20.5 Structural Notions in Linguistics

In the history of linguistics, various scholars have made a number of distinctions helpful in better understanding the manner in which language is organised and operates. The Swiss–French linguist Ferdinand de Saussure (1857–1913), who taught in Geneva at the end of the nineteenth and the beginning of the twentieth century, is particularly important in this respect. He did not write any linguistic books but after his death his pupils put together notes taken during his lectures into a book and published it posthumously as Cours de linguistique générale ‘A course in general linguistics’ in 1916. In the Cours, Saussure introduces a number of dichotomies which apply to all languages. Four of the most important are briefly described in the following.

Synchrony and diachrony: Synchrony is the investigation of language at one particular point in time (which may, but does not necessarily have to be, the present). Diachrony is the investigation of language over time and is really the investigation of several synchronic ‘slices’, ordered one after the other (Figure 20.2). Changes over time are hence due to factors operating in each successive synchronic stage, each ‘slice’ of a language.

Figure 20.2 Diachronic and synchronic views of language

Langue and parole: Saussure’s second major dichotomy is that between langue and parole. By langue is meant the system of the language; by parole is meant actually speaking a language. These two notions are close to those used by Noam Chomsky in the early 1960s, namely competence and performance. There is a difference between langue and competence, however. The former stresses the system of a language as the common core of linguistic knowledge in a community of speakers (here Saussure was under the influence of leading nineteenth-century sociologists like Émile Durkheim) while the latter refers to the abstract ability of speakers to produce and recognise correct sentences in their own native language on the basis of structures which they have abstracted during the period of language acquisition in their childhood.

Signifiant and signifié: For Saussure, a language is a sign system, which consists of two essential parts. The first is the signifiant (that which signifies: the sound shape of a word) and the second is the signifié (that which is signified: the concept in the mind). The concepts are usually linked to objects in the outside world, but that is not necessarily the case (Figure 20.3). For instance, the word unicorn is a signifiant which points to the concept, the signifié. But that is where it stops because there are no unicorns in reality.

Figure 20.3 Signifiant and signifié after Saussure

From this account it is clear that Saussure sees the concept as something mental, which is independent of language and of the external world. Take as an example the concept of ‘book’. For Saussure, we have a notion of ‘book’, which points to our mental concept of ‘book’, which then in a further step denotes an example of the physical object ‘book’. This last step can be present, as in the sentence The book I read last week (a particular book in the outside world); or not, as the case may be, such as Books are a key source of reliable information (a generic statement, no specific object reference).

The relationship of the word to the concept is arbitrary by which is meant there is no necessary connection between the sounds used and the concept being referred to. This type of relationship is what gives human language its specific quality as a communication system. Once language no longer reflected a word’s contents in its form (as with onomatopoeia, see the example of cuckoo in Section 21.1 below) it was free to develop new forms which needed only to be accepted (conventionalised) by the speech community. Each language consists of thousands of conventions which determine the sound shapes used for concepts (of objects or ideas). The arbitrary nature of these conventions can be seen in the variety of sound shapes used for one and the same concept, such as ‘tree’ which is arbre in French, Baum in German, derevo in Russian, dentro in Greek, crann in Irish, zuhaitz in Basque, etc. (see the comments in Relationship of words to concepts above).

Paradigm and syntagm: A syntagm is a series of linguistic units arranged horizontally (e.g. a phrase or sentence). Each syntagm consists of a number of ‘slots’ into which various elements can be placed. A paradigm is then the set of elements which can occupy a single slot in a syntagm (Figure 20.4).

Figure 20.4 Syntagm and paradigm

Apart from the four Saussurean dichotomies just discussed there are others, which are part of the bedrock of linguistic analysis.

The great German poet Johann von Goethe (1749–1832) once said that because people can speak they think they are entitled to speak about language.Footnote ¹ The point he was making is that because we have language we think we have the necessary knowledge to make pronouncements on language. However, this is not true. In order to make objective statements about the structure and use of language one needs training as a linguist. So why does one need technical vocabulary when talking about language? The reason is that, although we all speak our native language effortlessly, there is a lot of internal structure involved in this process and we are normally unaware of this. To describe all facets of language one needs an array of technical terms. Bear in mind that most of our knowledge of language is unconscious, like an iceberg where nine-tenths are hidden below the water’s surface. We internalised this knowledge in our early childhood without being aware of it during the language acquisition process. Now if you wish to discuss all the conscious and unconscious aspects of language, especially the latter, you need at least some technical vocabulary to do so. To help readers along, I will define terms as they are required and do my best to find everyday analogies to explain them.

21.1 How Words Represent Meaning

21.2 Linguistic Relativity

By and large it is true to say that languages have words for the objects of the world, the thoughts and feelings which its speakers have and experience. And to a certain extent it is the case that separate words for objects tend to reflect their relative importance for speakers. For instance, English has a special word for thumb, the finger on the inside of the hand at an angle to the others. But the equivalent on our feet, the big toe, does not have a special word.Footnote ⁴ One could say that one uses one’s thumb more and one sees it more often and so there is a separate word for it. But not all languages work like that. Indeed, some do not even have a separate word for ‘toes’; for example Irish uses a form méara coise ‘fingers of the foot’ (as does Turkish). In this case the pitfall is to imagine that the Irish (or the Turks) pay less attention to their toes because they do not have a separate word for them.

Here is a parallel example from the realm of thoughts and feelings. English has a word for ‘remember’ but there is no special word to express what happens when one suddenly has a feeling again which one had in one’s past, a type of emotional remembering. What speakers of course do is to use a phrase when a single word is not at their disposal, such as I remember the feeling I had on the morning of my wedding day. The danger here is imagining that the existence – or lack – of a special word for a particular matter somehow reflects its importance for the speakers of the language in question.Footnote ⁵

During the first half of the twentieth century many linguists in America were concerned with describing the remaining native American languages before they died out. There are several hundred of these languages belonging to a couple of large families and they are unrelated to languages in Europe and structurally very different from these. This fact led linguists to reflect on the possible influence which the structure of language has on thought. Two linguists in particular are associated with this idea: Edward Sapir (see Section 20.3 above) and his student, the anthropologist Benjamin Lee Whorf (1897–1941). The view that language substantially influences thought is known as the linguistic relativity hypothesis (formerly the Sapir–Whorf hypothesis).

This hypothesis has a strong version, namely that language determines thought and it is this which Whorf apparently adhered to. He assumed that the structure of Western languages is quite similar and termed this Standard Average European. The structure of the native American languages which he investigated, above all Hopi (a native American language of the Uto-Aztecan family, mainly spoken in north Arizona), is radically different from European languages. The proponents of the linguistic relativity hypothesis maintain that such structures in language influences the thinking of native speakers. Critics of such an extreme view state that distinctions which are not formally encoded in a language can always be expressed via paraphrase. A closer look at the native American languages, which Whorf originally considered, shows that they also have the option of paraphrase or have availed of other methods to overcome what they seemingly lacked. Hopi, for instance, has borrowed words from Spanish, e.g. kawáyo ‘horse’ from caballo; karéeta ‘waggon’ from carreta.

There is probably some validity to a weaker form of the linguistic relativity hypothesis. The structure of a language compels us to focus on certain aspects of what languages are used for and it can also highlight certain features of extra-linguistic reality. Languages compartmentalise our experience of the world in different ways and segment the continua of our lives. They play an essential role in organising the world we experience into categories which we can handle with cognitive ease.

21.3 Language as a Reflection of Speakers’ World

Exobeings would live in the same three-dimensional universe as we do. They would be subject to the same arrow of time which we perceive on our Earth (you cannot be ‘unborn’, a broken glass cannot be ‘unbroken’). They would furthermore have a conception of actions in time. How could they not have such notions, such as to run, turn, jump, slip, etc.? They would have notions of objects, their sun, their planets, possibly a moon or moons. They would have a notion of subject and object, that to eat something or to be eaten by it are two very different things. Some words (our nouns) could be animate, like themselves or other animals, and some inanimate, like a rock or a stick. They would perceive differences in size, such as big rocks and small rocks; angry animals and peaceful animals. And they would recognise different spatial relationships in their world, the animal is under the rock, in the tree, beside the river, on the strand, etc. Furthermore, they would recognise differences in the manner in which actions are performed, to run quickly (escaping a predator), to walk slowly (creeping up on prey). And exobeings would make decisions by weighing up the factors in a given situation, for instance when considering ‘fight or flight’ on being confronted by aggressive animals.

All these cognitive distinctions would result from the environment in which they live just as they have in our environment on Earth. It is reasonable to assume that an exolanguage, with a similar level of differentiation to ours, would have means to express these distinctions. So how would this work? Consider how it is done in human languages (Table 21.1).

With this very simple arrangement we already have five major word classes. Another extension would be to use spatial elements for temporal relations; for example, consider the following two sentences, where the first meaning of near is spatial and the second temporal: (i) The house is near the sea; (ii) The election was near Easter. Another extension is where literal meanings become figurative – a strong tendency among languages – consider (i) The forest is beyond the river and (ii) This is beyond our capacities at the moment.

Apart from the categories listed in Table 21.1, languages generally have conjunctions which serve to link parts of sentences which often stand in a certain kind of relationship to each other, such as: Fiona likes astronomy and physics (and links two elements), Fiona is studying astronomy although she prefers physics (although is a concessive conjunction). Furthermore, there are languages where relations, such as those involving space or time, are expressed via endings on words; this depends on the ‘type’ of a language, a term referring to the overall structural organisation of a language. In Finnish, for instance, the relations shown by prepositions in English are normally expressed by endings; for example, He tapasivat talossa ‘They met in the house’ where -ssa ‘in, inside’ is an ending added to the word talo ‘house’.

21.4 Names and Language

All languages distinguish classes of objects and individual instances of such classes. For instance, there are dogs and cats – classes of animals – and there are individual animals – single dogs and cats. People use names for individuals to which they have a specific connection, such as Fido for your pet dog or Sandy for your pet cat. Names are used to reference persons, so much so that all human beings have names and these are a key part of personal identity. Complex multi-part names are often used to distinguish people from each other, especially because it is common in cultures for names to be selected from a small set. Just think of how many males are called John or Michael and how many females are called Mary or Anne (in countries with a Christian tradition).

All languages have names too. The name used for a language may derive from its community or be that used by outsiders. A name may refer to a region of origin, as with ‘English’, which means settlers from Angeln, an area in the far north of Germany, close to the Danish border. A name may simply derive from something like ‘we, of the people’ as with German deutsch and the related Dutch word duits or Swedish tysk. The etymology (antecedents of a word in history and its meanings) can be traced back to Proto-Indo-European *teuta ‘tribe’ (note that the asterisk in this case indicates a very early form, which is postulated but not actually attested).

Objects tend to get names when people devote increased attention to them. All the planets in our Solar System have names. Those outside our system are referenced using a numbering system, often with a leading name tag such as the Gliese set of exoplanets, named after the German astronomer Wilhelm Gliese. There is also a group of exoplanets named after the Kepler Space Telescope, which was instrumental in finding them. For stars and constellations, mythological figures from classical Greece and Rome have served as a good source of names, usually with a certain semantic motivation, so that Jupiter, the largest planet, is called after the chief Roman god; Mars, the red planet, after the god of war; Pluto, the furthest planet (now demoted from this status), after the god of the underworld.

21.5 Language, Environment and Culture

Languages reflect the environments and cultures in which they are embedded. The more general of these two situations concerns environment as can be seen when one considers the terms used for basic colours across the world. This phenomenon was investigated in a well-known study by two American linguists, Brent Berlin and Paul Kay, in a book published in 1969 called Basic Color Terms: Their Universality and Evolution. They suggested that languages can exhibit seven levels of colour distinction, starting with dark-cool (approximately ‘black’) and light-warm (approximately ‘white’) proceeding through red, followed by either green or yellow, then blue, brown and, finally, other minor colours such as pink, purple, grey, orange. The first universal level, with ‘black’ and ‘white’ archetypes, rests on the division of night and day; the next level ‘red’ most likely has to do with the colour of blood and the general association of red with vitality. The colours green and yellow reflect common colours in nature, etc. From their investigation Berlin and Kay concluded that there are 11 basic colours, which occur in a specific implicational order, for instance the authors claimed that if a language had a word for ‘green’ then it had a word for ‘red’, if ‘brown’ then ‘blue’, etc. The ordering for their eleven colours is as shown in Table 21.2.

The basic colours found on an exoplanet would presumably have a similar motivation: the environment in which exobeings live would determine the colour distinctions their languages would make.

21.6 What Do Speakers Know about Language?

When non-linguists think of knowledge they think of conscious knowledge, that is, of something which they can express and reflect on. For instance, if someone asks you whether you know how to play chess, you would answer ‘Yes’ if you felt able to list the rules of the game. Knowledge of chess would be seen as equivalent to the ability to use the pieces and explain the rules governing their movements to others. This is a typical instance of conscious knowledge as players are aware of the rules and reflect consciously on them during the game.

Language is organised quite differently. The distinction between unconscious and conscious language not only refers to what words and structures are well formed but also to what is not possible in a language. Imagine you had to invent a name (in English) for a new brand of dog food and someone suggested the word fnoppy to you. It is unlikely that you would accept it because you know (unconsciously) that in English /fn-/ is not a permissible beginning for a word (technically a syllable onset), although it existed in Old English (450–1066 CE), as in fneosan ‘to sneeze’, and is allowed in other languages like Russian (as vn-, e.g. vnimaniye ‘attention’). In fact, you know (again unconsciously) that only a fricative (a sound produced without interrupting the air flow fully) produced at the same point in the mouth can occupy the slot before a nasal sound here: snoppy would be a permissible sequence but thnoppy or shnoppy would not, as neither th- nor sh- is produced at the same point as n. Now the fact that English speakers cannot formulate the restriction on word beginnings (syllable onsets) simply means that this knowledge is unconscious. It cannot be verbalised by non-linguists. But that does not make it any less valid or true.

Take another example, this time from grammar (syntax). The following sentence is unlikely to be accepted by speakers of English: *Saw I down on the strand her. The reason is simply that they know that in statements the subject precedes the verb, so we have I saw …, and that the direct object always precedes a prepositional phrase, as in … her down on the strand. You would never think of producing such a sentence because when speaking you automatically avoid structures which are ill-formed in English. However, the order of sentence elements varies greatly across languages. For instance, in Irish the order of elements in the ill-formed sentence of English just quoted is quite normal, as can be seen from the Irish equivalent of this sentence: Chonaic mé shíos ar an trá í, literally ‘saw I down on the strand her’. Such examples show that there is variation in word order across languages, that speakers recognise this and that they unconsciously know (from language acquisition in early childhood) what the order is for their native language, that is what order will result in well-formed sentences acceptable to other native speakers. So people do not think of the rules of syntax when they speak. It is valid to assume that there are many rules governing well-formedness in sentence structure; however, these rules cannot normally be listed by native speakers (unlike the rules of chess) unless the speakers in question have received specialist training in linguistics.

Unconscious knowledge must exist, otherwise our speech would be incomprehensible. Here is a comparison to illustrate what is meant: if you think about how you walk, then you probably imagine yourself putting one foot forward, shifting your body weight onto this and then moving the other foot forward, then shifting to this other foot and so on. This would be the equivalent of conscious knowledge of language. But walking requires a very intricate interplay of nerves, muscles and tendons as well as feedback to the organ in the head responsible for balance to ensure that one remains upright during the action. The neurological and muscular activity involved is not ‘visible’ to the person walking and is similar to the unconscious knowledge of language which is active ‘in the background’ when we speak. It is right for the human organism to keep this knowledge hidden from speakers as too much consciousness would render speech too difficult – the main thing is that the unconscious knowledge works properly, which it does nearly all the time. The only exceptions being unusual situations, temporary ones such as nervousness, tiredness or inebriation, when one’s speech is somewhat uncoordinated. There are also more lasting disturbances, known from language pathology, which can arise after an accident or brain disease such as a tumour or a stroke.

Our unconscious knowledge about language is in part innate – the universals of language which we inherit in the genetic code of our DNA – and in part it is acquired in the early years of childhood. Such unconscious knowledge works very well, which is why a language that is acquired early is entrenched for life. After puberty the ability to acquire knowledge with comparable competence declines rapidly, which is why adults have difficulty in learning a foreign language.

In the evolution of our language faculty various elements of language were relegated to the unconscious. Several advantages accrued from this development. One is speed: unconscious actions are always faster than conscious ones, just think of how quickly you can tie your shoelaces if you do not think about it. Another advantage is that it frees up conscious resources for more important aspects of language. Consider that you do not make a conscious decision for each item in language. Imagine deciding consciously to pronounce /f/ then /u:/ then /d/ when saying the word food. And, when you say a sentence, you do not consciously choose a syntactic structure and decide what precise elements to put into it. When saying something like What is the food like in that hotel? you think of the meaning you wish to convey and the well-formed sentence is generated unconsciously. On reflection you realise that, if you are a native speaker of English, you certainly will not say something like What that in hotel like food is the? This brings us to the issue of intuitions about one’s native language.

21.7 What Are Speaker Intuitions?

If linguists are not sure what structures are valid in a language, they may choose to interview a representative selection of native speakers on this issue. For instance, some verbs in English take an infinitive and others take a participle (technically called a complement). To determine what options are valid for a particular verb one could elicit responses from test persons by presenting them with templates like the following and asking them to fill in the empty slot.

• Fiona considered ____ homeFiona wanted ____ home

The answers would probably be as indicated below (a tick shows the acceptable and an asterisk the unacceptable sentences to speakers of standard present-day English; a question mark indicates that speakers might be a bit doubtful).

• ✓Fiona considered going home✓Fiona wanted to go home

• *Fiona considered to go home? Fiona wanted going home

There are a number of important generalisations one can make from the preceding sections. The first is that in the evolution of the human language faculty certain elements migrated from conscious to unconscious knowledge, which freed up capacity for conscious reflection on meaning in language production. The second is that the knowledge of our native language, which is stored unconsciously during the first language acquisition process in early childhood, has the effect of yielding a large amount of agreement among speakers about what structures are well formed in their language and what ones are not. This in turn has made languages more efficient as communication systems in speech communities, a quality which they retain to the present day.

Linguistics is the study of the human language faculty and the languages it engenders. It enjoys interfaces with other sciences such as cognitive psychology, neurobiology and physiology. The borders between linguistics and other sciences are fluid and have shifted with increasing research in the field. Where we draw the lines is a matter of debate among scientists, but we do draw them because science compartmentalises reality for the purpose of inquiry and analysis.Footnote ¹

For some two centuries linguistics has been an established branch of science, and considerable strides have been made in researching how humans speak, how they acquire their native language and in documenting the known languages of the world. This knowledge has been arrived at by devising means for analysing human language and by adopting approaches based on notions which are likely to be quite different from the views which non-linguists will have about the subject. Thus, one of the difficulties for linguists in their attempts to explain the nature of language is that non-linguists will always have ideas about language already. These derive from reflections on the more conscious parts of language such as vocabulary; on the widely held belief that language change is language decay; on ideas about what constitutes good and bad language, in particular in connection with style and social class, and on an inordinate reliance on the written word; and, finally, on a general confusion of language itself with the people who use language.

In all these areas linguists will find that the ideas of non-linguists need to be adjusted and set in a new direction. For example, linguistics is often more concerned with the less conscious parts of language, such as the levels of sound systems and sentence structure, and it views language change as an inherent and necessary aspect of language, seeing as how it has its ultimate roots in the continual choices made by humans when speaking. The goal of linguistics is not only to describe languages but language as a whole and to reach conclusions about our unique ability to speak, that is, the human language faculty. Linguistics is certainly not concerned with dictating usage to others, called prescriptivism, and it is primarily oriented towards the spoken word, the spelling system of languages being a secondary phenomenon arising much later in history due to needs in societies.

22.1 The Complexity Envelope of Language

The complexity envelope of human languages is an issue with two different but related aspects, a physical and a cognitive one. The physical envelope is defined by the sounds which humans can produce with their organs of speech (see Section 22.4 below). The cognitive envelope is determined by our mental capabilities regarding language. There is no easy metric for measuring these, but we know that the structural similarities across human languages are ultimately determined by our cognition. Looking at the documented languages of the world one sees that there are great differences in relative complexity of their subsystems.Footnote ² Take sound structure, for instance. All languages have a set of key sounds used to distinguish meanings when forming words, thus p and b are used to distinguish word pairs like pat and bat, pear and bear (technically such sounds are called phonemes). The discussion here concerns sounds, not letters, so a word pair like cease and seize, while written very differently, only have one sound difference, in the ‘hard’ or ‘soft’ s at the end.

The numbers of these key sounds (phonemes) varies greatly. English, with 40+ (the exact number depends on certain interpretations and on the variety of English one is considering), is somewhere in the middle. But there are languages with fewer phonemes, such as Hawaiian with only 13 (8 consonants and 5 vowels) while other languages, like those found in the Caucasus (the mountainous region between the Black Sea and the Caspian Sea) can have as many as 70–80. The lower limit is probably determined by the number necessary to provide sound shapes for the words of a language. It would be very difficult to attain different sound shapes for thousands of words if a language only had three phonemes, for instance. The upper limit for phonemes is probably determined (i) by processing difficulties with huge numbers of phonemes and (ii) by the limits in sound variation which a human can realise. Furthermore, in languages with very high numbers of phonemes, usually only some occur in certain contexts, at the beginnings or ends of words, only in the middle, only in certain clusters, etc. This again facilitates the processing of sounds by speakers and reduces the cognitive burden of handling a very large number of phonemes.

Languages also vary considerably in their grammar, specifically in the part of grammar concerned with cases (nominative, accusative, genitive, etc.), number (singular, plural), gender (masculine, feminine, neuter), tense marking (present, past and future) and the formation of words. It is well known that German and Russian have complicated grammars with three genders (masculine, feminine and neuter) and lots of endings. Finnish and Hungarian have very complicated case systems but no grammatical gender.

There are many views on why some languages should show greater complexity in grammar than others. Those which have little, such as English or East Asian languages like Chinese languages or Vietnamese, get on perfectly well with few endings and may have other mechanisms for making key distinctions in language, for example by using different tones in words, as in the different languages in China. But why should a language like English be grammatically less complicated than German, to which it is related, or Bulgarian compared to Russian, to which it is also related? One reason might be that contact with other languages leads in time to a simplification of the grammar because there are many second language speakers in the community and they use simplified grammar. The converse would seem to be true: languages which have long existed in isolation tend to have complex grammars, like those isolated in the Caucasian mountains for centuries on end.

22.2 Levels of Language: Modular Organisation

Language is a phenomenon which evolved over tens of thousands of years and can be compared with parts of the human organism more than with man-made structures. In order to maximise efficiency and minimise the effect of damage to some part, language evolved into a system which is modular in its organisation (for more information see Section 32.7 below). In this respect it can be compared to a part of human physical make-up, like the immune system (see Section 32.8 for further discussion). Each module of language is self-contained with its own rules and representations – for instance, phonology (sound structure) and syntax (sentence structure) have quite different internal structures. There are, of course, interfaces between each module so that they can interact together and appear as a whole in actual speech. What unifies all modules is the purpose of the system. Just as the immune system has the superordinate goal of protecting the body from infection, language has the function of organising and expressing thoughts, thus enabling communication: ultimately, all elements of language work towards this two-pronged end. However, because language is a cognitive ability of humans it also has secondary functions. For instance, it has a frequent role as a carrier of social attitudes, something which has nothing to do with the simple communication of messages but which has been superimposed on the system. This is because speakers have always forms of language associated with those who use it and are aware that the system can be used to differentiate speakers socially without a loss of the primary function.

It is common procedure to treat the various levels of language (Figure 22.1) separately, as in textbooks on linguistics. This has the tuitional advantage that one can deal with them concisely and neatly in separate sittings of a course. However, one should emphasise that the division is not something people are usually aware of when speaking. Because of this they do not always expect language in linguistics to be divided into levels. Arguments in favour of the psychological reality of the different levels can be put forward, for example by showing that the rules of phonology are quite separate from those of syntax despite the interface which exists between the two levels. In the following, the different levels are dealt with from the bottom up, as shown in Table 22.1.

Figure 22.1 Levels of language

22.3 Language Typology

Language typology attempts to classify languages according to higher-order principles of grammar (morphology and syntax) and to reach generalisations across different languages, irrespective of their genetic affiliations, that is, what language family they belong to. It is also concerned with making statements that apply to all or nearly all languages, which are called universals. The number of universals is actually quite small (but essential to all languages), so language typology is often concerned with determining what the most common value for a parameter is across languages. Take word order, for instance. All languages have what is called a default value for word order, that used in simple declarative sentences. In English the default value for word order is subject–verb–object (S–V–O) though this may vary, for example in interrogative sentences, as in Can we leave now? Some speakers will accept as well formed sentences which have the order O–S–V, as in This girl I don’t know, though such word orders are not really part of standard English.

Some default word orders are frequent across the world’s languages. S–O–V, with the verb at the end, is found in languages like Turkish, Basque and Japanese, though S–V–O is the most common word order when languages do not have many grammatical endings to distinguish subjects from objects. Other word orders, like V–S–O (verb-initial languages) are much less common, Irish and Welsh being two notable examples in the European arena, Classical Arabic and Tagalog (a native language of the Philippines) are other examples.

The demographic situation of a language can play a role in what categories are present. For instance, languages which have been in geographical isolation for many centuries tend to develop certain categories, as opposed to languages in scenarios of high contact with other languages. For instance, verb forms which denote evidence for a statement tend to occur only in small, close-knit communities, where indicating responsibility – or lack of it – is expected of speakers who are known to each other as in Eastern Pomo (a Native American language in Northern California).

22.4 Language Production

Humans have very controlled breathing (when at rest or just walking), with about 10 per cent of the time devoted to inhaling air and the remaining 90 per cent to exhaling. It is during the period of exhalation that we produce speech. This is an unconscious process and we do not perceive of breathing in as gulping air.

By comparison, when we run, we usually coordinate our breathing with the rhythm of our running and so cannot devote 90 per cent of the time to exhaling and speaking. The intake and release of air takes about the same amount of time, which makes it difficult to run and speak at the same time.Footnote ³

Bipedalism probably assisted greater breath control among early Homo species, as the movement of the legs and the rhythm of breathing were no longer as closely connected as with quadrupedal animals. With bipedalism, the forelimbs were not primarily involved in movement in our surroundings and the further evolution of our forelimbs to arms with hands having fingers, with an opposable thumb, provided us with great dexterity for additional tasks apart from basic locomotion. These developments further disconnected breathing and the use of our limbs.

22.5 The Human Tongue and Throat

Babies when born have their larynx, the ‘voicebox’, where the vocal folds are located, quite high in the throat, in fact so much so that they can breathe and swallow food at the same time, as can many animals. But after about three months the larynx descends into the throat, producing a narrow, inverted fork-like configuration which allows us to produce sounds with our vocal folds quite far down in our throats, and we have a resonance chamber above it to amplify the sound we are producing. This dual-purpose configuration of food plus windpipe makes us prone to choking, as food can easily enter the trachea (windpipe) if it is not closed off by the epiglottis (Figure 22.2).

Figure 22.2 Human breathing apparatus

But another advantage of the lowering of the larynx, for us as a species and for each of us individually, is that it allows a lowering of the tongue root, which in turn allows the tongue to move easily along a front-to-back (horizontal) axis, not just an up and down (vertical) axis in the mouth. This means that the human tongue after early childhood can adopt any of the positions for vowels which might be necessary for any language.

Much has been made of the lowered larynx in early humans as possible evidence of the ability to produce speech (Figure 22.3). But research by various scholars, such as W. Tecumseh Fitch (Reference Fitch and RebyFitch and Reby 2001), has pointed out that a lowered larynx is regular with other mammals, like dogs when barking or stags when roaring. But these animals appear to lower their larynx only when producing their characteristic sounds whereas humans have the larynx permanently lowered. Nonetheless, the rapid development of the primate larynx is disproportionate to the development of body size with these animals, not just humans (Bowling et al. 2020).

Figure 22.3 Human larynx (anterior and sagittal view)

The crucial distinction between mammals and humans lies not so much in anatomical differences but the neuronal circuitry controlling the vocal organs. This can be seen, for instance, in the size of the hypoglossal nerve canal, which in turn indicates a sophisticated nerve bundle, innervating a flexible and agile tongue necessary for speech (Reference 652HurfordHurford 2014: 85). The same is true of the nerve canals which innervate the chest muscles and the region of the throat.

22.6 What We Hear

Would exobeings be able to hear? Clearly, they would need to receive acoustic input from their surroundings and from others. To process spoken language, one must be able to hear. For the latter, one needs ears or some functional equivalent. Higher mammals have ears and can generally hear as well as we can, or even better, in many cases. So they have one of the preconditions for successful (spoken) language. Our hearing is especially attuned to sounds between about 20 hertz and 15,000 hertz (less at the high end as we get older), which is the range of human speech. Non-linear sounds, big jumps in frequency, very high-pitched tones, very low booming sounds, trigger the startle reflex, before your conscious brain knows it. This is a primitive reflex which prevents us from being attacked and possibly eaten by some nearby predator.Footnote ⁴

The range of sounds for human speech is also that of noises in nature, such as rushing air, moving water, leaves brushing up against each other or of some animal walking in the undergrowth, all noises the perception of which were essential to species survival during our long evolution. These sounds would be found in a similar wavelength range on an exoplanet so it would make sense to have hearing which is attuned to such sounds and to have the sounds of an exolanguage located within that range.

Our hearing is especially sensitive to the sounds of language in that we can recognise language in any set of noises we happen to be picking up at any given time. This ability would seem to be due to the processing of incoming audio signals in the primary auditory cortex of our brains. It is here that we can allocate sounds to the categories in the system of the language we are listening to. This is no mean feat and involves normalising the input signal, which means taking into account the relative frequency of a voice (male or female, child or adult) and the tone of voice of the speaker. And this is done in real time while also deciphering the grammar and meaning of what one is hearing.

22.7 Vowels and Consonants

All languages have a distinction between vowels and consonants. The vowels are produced (i) by allowing the vocal folds in the larynx to vibrate in the escaping air from the lungs and (ii) by adopting a certain configuration of the oral cavity (the space in the mouth) by moving the tongue into a certain position (Figure 22.4). Consonants on the other hand are all produced by obstructing the escaping air in some way or other. There are consonants which are vowel-like, involving little obstruction, such as <r> and <l> in rip and lip, respectively, and others which not only involve a more radical obstruction but also do not require any vibration of the vocal cords, such as <t> or <s> in tip and sip, respectively, or the <(t)ch> sound in a word like chip or catch.

Figure 22.4 Vocal tract with points of articulation

The rise of consonants in early human language is not surprising given that the tongue, lips and velum are part of the oral tract, which are moved about repeatedly when we are biting, chewing and swallowing food. The human tongue was agile enough at the beginning of the human lineage to produce primitive sounds, which became increasingly sophisticated in their articulation with the continued evolution of humans and their speech.

How Are Vowels Produced?

Imagine you go to a concert. Along with all the others you take your seat, the members of the orchestra come onto the podium and sit down. When everyone has settled, the chief violinist stands up and gives a sign to the oboe player, who then plays a single long note, an A, and the rest of the orchestra chime in by playing the same note, aligning the pitch of their instruments to that of the oboe. An A is about 440 hertz, though some orchestras have a slightly higher value (a few hertz more) to create a more brilliant sound. But the question is: if all the instruments are playing the same note why do not they all sound the same? The answer lies in the harmonics produced by each instrument. The sound of an instrument consists of a note at a certain pitch but also of various frequencies above that note which have stronger or weaker values, depending on the type of instrument (wind or string, for example) and its shape (trumpet, horn or tuba, for instance). So the vibrations of the air when an instrument is being played consists of the basic pitch of notes, called the fundamental frequency, and a highly complex pattern of additional vibrations above that pitch, called overtones, which results in the typical sound of an instrument. So the pattern of overtones produced by an oboe, a clarinet, a flute or a horn are all very different.Footnote ⁵

So what does this have to do with language? It explains why the vowels of a language all sound different. Imagine you produce the vowel of the (English) letters ‘AH’ (as in the word shah), long and clear. Now keep the pitch you are using for that vowel in your throat (produced with your vocal folds) and produce the sounds for the (English) letter ‘E’ (as in the word see) and then the letters ‘OO’ (as in the word soon). So now one could ask: if the pitch is the same for all three vowels, why do the AH, E, OOFootnote ⁶ vowels all sound different? The answer lies in the variable amplitude of the overtones, which in phonetics are called formants. Formants are derived from the different configurations of the tongue (Figure 22.5) and hence the difference in the shape of the oral cavity through which the sound generated by the vocal folds passes, on its way out through the mouth.

Figure 22.5 Vowel quadrangle

Making use of what nature gives us

The air from the vocal folds normally escapes through the mouth when speaking. But it is also possible to open the nasal cavity at the same time by lowering the velum at the back of the mouth. Some individuals lower their velum slightly during vowels in English, this is then just an idiosyncrasy of their speech. But if a whole community does this consistently, the language may develop nasalised vowels as happened in French, cf. quand [kã] ‘when’ (historically, the nasal quality of the <n> spread to the left into the preceding vowel and the [n] sound was lost; Polish also developed nasalised vowels in a similar fashion). Such extensions to the palette of possible sounds could be expected in exolanguages as well: anatomical features which would increase the differentiation among sounds would likely be co-opted into the sound systems of at least some of the exolanguages of a given exoplanet.

22.8 Convergent Evolution and Language Production

For scientists there is a problem with something which only occurs once. Is it a fluke occurrence or a common development, given the right conditions? Humans have a set of folds in their throat which vibrate to produce sound. Was this development unique or are there independent parallels elsewhere in the animal world, which would support the idea that this anatomical feature was not just a freak of nature? If one looks at some other mammals, in this case toothed whales and dolphins (odontocetes), one sees that they have phonic lips in the part of their anatomy which corresponds to the nasal cavity in humans. When the phonic lips are brought together during exhalation by the animal, the ambient tissue vibrates producing sound analogous to that generated by humans with their vocal folds (in the larynx).

This is a clear example of a convergent development, which shows that the production of sound for echo location in odontocetes and that for voice with humans avails of similar anatomical structures, supporting the view that these are naturally evolved mechanisms for sound generation.

23.1 The Nature of Language Acquisition

The language faculty, the ability to understand and acquire human language, is a feature of our neurobiological make-up which has been passed down through the generations in every human being as part of our genetic endowment.Footnote ¹ It is the language faculty which allows us to acquire any language as long as we are exposed to it in our early childhood. Although it cannot be directly observed, the language faculty imposes structural conditions which must be met by all the languages of the world, that is, it provides limits to what can occur in a human language by containing a framework within which language variation can arise.

The language faculty is the same for everybody no matter where they might come from. This can be shown quite easily by the fact that a child from one part of the world, if taken to another, at an early enough age and exposed to a language there, will acquire that language as a native speaker, indistinguishable from all the others whose parents are from that part of the world. For instance, if a British couple moved to Vietnam with a newborn baby and stayed there for some years (to allow language acquisition to take its course through interaction with other Vietnamese), then the child would learn VietnameseFootnote ² like any child of Vietnamese parents.

It is clear from the above remarks and from basic reflection that we have the ability to acquire and understand any language, but that we are also native speakers of one or maybe two, rarely three, of the over 6,000 languages spoken on Earth. These are all external languages, by which is meant the expression of the internal language faculty, manifest in pronunciation, grammar and vocabulary when we speak.

The framework of constraints and conditions which the language faculty puts on externalised languages has been labelled ‘universal grammar’ (see above) and is manifest in the similarities in underlying structure of all languages. Knowledge of this structure is passed on genetically to each generation and in our long childhoods we take this abstract structural knowledge and combine it with the specifics of the language or languages we are exposed to during the process of language acquisition. Just what features constitute this universal grammar is contested among linguists but the existence of the framework, as evidenced by the deep structural similarities across all languages, is far less disputed.

The extended childhood of humans allows them to develop their faculties, such as language, more fully than if they reached adulthood more quickly.Footnote ³ There is also a sense in which humans are born too early (from a purely physical point of view). A possible reason for this could be to allow the child to pass through the mother’s birth canal with its large head. A consequence of this is that the child enters the outside world quite early and can receive and process language and other stimuli from this source at an early point in its physical development.

23.2 The Question of Modality: Sound or Gestures?

Most animals use sounds in the audible range to communicate with each other. It is true there are bats which use ultrasound (above 20,000 hertz) to locate their prey. One might think that exobeings could also use ultrasound for their communication and so their language would be inaudible to us. But how likely is this? In our animal world, ultrasound is the exception rather than the rule, although many more animals have been discovered to use ultrasound, such as tarsiers, and that might be true for exobeings as well. But consider that ultrasound (used in echo location) is mostly typical of nocturnal animals, especially those which are insectivorous. Of course, if you had beings on the dark side (or twilight zone) of a tidally locked exoplanet, they might avail of ultrasound.

Compare this situation to temperature range: most animals live in a range which does not exceed 40 °C. But there are extremophiles which can live around or indeed above boiling point. However, that is very much the exception in our world. So maybe sounds of the audible range we use, like the normal temperature range for carbon-based biology, are the most typical in the universe and so the most likely to be used by exobeings.

However, the fact that sound is most likely as a modality does not mean that it is the only one (see following table). Recent research has also shown that the modality for expressing internal language can be either by vocal–auditory means (speaking and hearing) or by signing (using one’s fingers and hands, possibly with facial expression and mouthing, in precisely structured movements for others to see and interpret).

Reference Goldin-Meadow, Tallerman and GibsonGoldin-Meadow (2012) found out that young children have two-part communication strategies to begin with: they point to something and utter a word and later, when the two-word stage develops, they drop the gesture. If this is a case of ontogeny recapitulating phylogeny (the growing individual going through similar phases as in evolution), it might be evidence that gestures preceded sound for language expression in the history of our species.

23.3 Sign Language

There are advantages to sound: it can be heard over several metres, one does not need to be facing the source to hear it and it can be heard in the dark. It fades rapidly so that a reasonable amount of information can be packaged in a given stretch of time. Sound can be varied in various ways, for example by loudness, pitch and frequency, so that language can utilise these features, either for the basic distinction of words and for carrying additional information relevant in a discourse. This means that in evolutionary terms sound may well have superseded gestures, given its advantages. However, in situations where sound it not available, for instance with congenitally deaf people, gestures can and have been used. Gestures and language are closely linked so that it is no surprise that sign language arose as an alternative modality. Over the past few centuries flexible and sophisticated sign languages have been constructed, which act as a viable alternative to language in sound.

Sign languages have been developed for most of the large countries of the Western world, such as British Sign Language and American Sign Language. These languages are not necessarily related to each other (though American Sign Language and French Sign Language are), nor do they derive from the spoken languages of the countries where they are found. In the English-speaking world alone, there is a plethora of sign languages which have arisen in the past few centuries or so: South African Sign Language, Australian Sign Language, New Zealand Sign Language, Irish Sign Language, etc. (these languages share features with British Sign Language, less so with American Sign Language).

Of particular interest to linguists have been the cases of new sign languages arising spontaneously in communities which have been relatively isolated from the speech communities around them and/or which have a high incidence of inherited deafness. For linguists, the essential question is how structure and conventions arise in such recent sign languages. There are four well-known instances.

In these cases it has been observed that the languages moved rapidly from an initial uncoordinated stage of idiosyncratic signs to a more conventionalised set of signs exhibiting an internal structure in the composition of multi-word phrases (corresponding to syntax in spoken language). The British linguist Simon Kirby has shown in experiments that this can happen quite quickly and that users abide by the conventions established in the group, thus enabling a transmission chain across future generations (see ‘Structure and Iterated Learning’ in Section 32.5).

The generalisation to be made from such observations is that human language may well have evolved in this manner itself at its earliest stages, with the conventionalisation of signs and the fixing of structures providing a consensus framework in a community for the use and transmission of structured language.

23.4 Communication by Touch?

Imagine you wished to communicate with someone you knew who was deaf and you did not wish to use sign language for some reason. You could devise a system whereby you tapped various rhythms with different fingers on the forearm of that person and vice versa. You would need to come to an agreement about what sequences meant and you could slowly build up a small vocabulary and some basic syntax, with simple sentences like actor–action–goal – as in I made some coffee – or some simple questions like Would you like some coffee?, all worked out via a system of taps on the forearm.

Now imagine a situation where some infants were born both aurally and visually challenged and they were encouraged by their caretakers to tap each other on the forearm to communicate with each other. The question is whether these individuals would develop a fully functional language consisting of taps on the forearm? The proponents of the view that internal language can be externalised (expressed) through a variety of modalities would say that this is in principle possible. It would not be very efficient, the tapping would be relatively slow, would not allow for the amount of structural variation of human speech, or signing for that matter, and would be cumbersome, as you would need to be close to your interlocutors and their forearms would have to be bare (not very practical in cold climates). But in principle it should be possible, it would be like playing the piano with one hand on someone’s forearm (note that there are instances where signers use touch to communicate with others, as in situations of darkness).

23.5 Receptive Modality

We normally listen to sounds and decode the semantics of the message in real time without any effort. But we can also scan written language, typically on a page, effortlessly composing the meaning the author of the text intended, again in real time. And the medium can vary: written text can be handwritten, printed in various fonts and size, upright or slanting, on paper of different colours, on metal, stone, wood, etc. or on a computer, tablet or smartphone.

When reading we scan the beginning of each word and, by means of an internal predictive algorithm, recognise words immediately and so can skim from word to word very quickly. We also switch from the end of one line to the beginning of the next without noticing it. Another feature of this visual recognition of speech in writing is that the image we produce in our visual cortex, from the input the retinas of our eyes pass back along the optic nerve, is stable whereas it should in fact show drag as our eyes move along each line of writing and then jump back and down to the beginning for each new line. Why do not we see the drag of eyes moving as we scan the lines on a page? The answer is that the image we see is one which is constructed in our visual cortex from retinal input. And this input is also scanned linguistically to allow us to understand the language which is printed in front of us. It is the processed image of the page and the linguistic interpretation which is presented to our consciousness. All of this happens effortlessly and seamlessly as we read text on a page. So where, you might ask, is this processed image in the brain? It is not really at any one point. We know that, when we are reading, both the visual cortex and Wernicke’s area (responsible largely for understanding language) are active. But information is also passed to and fro between these areas and between the visual cortex at the back of the head and the prefrontal cortex, which is responsible for most cognitive processing in our brain.

The net result of all this brain activity is that we perceive the page we are looking at as a three-dimensional image, projected outwards into space in front of us. This ‘illusion’ is created by the brain on the basis of retinal input passed to it through the optic nerve. Experience has shown us that the images we process from this input seem to correspond well to reality. After all, you might say, we can reach out to the paper with the print on it, we can measure the printed lines, count the number of words on a page, etc. But all this information is gained through a similar complex interaction of sensory input to the brain – from our eyes and our limbs (probably hands/fingers in this case).Footnote ⁴

If the image we think we see is really constructed in our brain, what do the eyes actually see? Well, light does enter the eyes and strike the retinas, leading to electrical signals being generated which are then bundled and passed along the optic nerve to the visual cortex. But the light passing through the lens of the eyes is flipped so that it is projected upside-down on the retina. Furthermore, although the image you see when you look out at the world seems to be in focus, your eyes actually only focus on a small part in the centre of your field of vision. Try the following little experiment to demonstrate this. Pick any line of text on the page you are now reading. Focus intensely on a single word in the middle of this line. And without moving your eyes see if you can recognise the words to the left and right of the one you are focusing on. You probably can only recognise one word to the left and right, two if you are lucky, but that’s all. Of course, under normal circumstances, if you wish to read words elsewhere on the line or page, you just move your focus to those words, like when you reread a line or skip ahead. But the point is clear: all you can see, in the sense of focus on, is a tiny patch in the centre of your field of vision. Your visual faculty is not like a camera where everything which is not in focus is blurred. Instead, it creates the illusion that the entire field of vision is in focus. But what in fact you are doing is endlessly and seamlessly shifting the centre of focus around your field of vision to take in information from what is in front of you. And you do this all day long, not just when you are reading as you are now.

From this simple example of how we read text on a page you can recognise that the brain is a highly complex, indeed manipulative organ, which constructs consciousness inside your head allowing you to manoeuvre through the outside world and to take in and utilise information as you do so. It is the seamless manner in which the brain processes the sensory input it receives that gives us the feeling that what we perceive is exactly what is outside in the real world. In our everyday lives,Footnote ⁵ there is no time when the brain flashes a message, ‘Processing, please wait’, with a little egg timer across our eyes, so we do not notice what it is going on below the level of our consciousness.

23.6 Language and Writing

We are not pre-wired to write languages, but in those countries with written forms of language (most) children learn to write from an early age. The key here is the specification ‘from an early age’. If children learn to write before the age of four or five, they master the writing system of the language in question. There are a number of hurdles in achieving competence in writing. The children have to master the principles of the writing system.

23.7 Linguistic Diversity on Earth and Beyond

The human language faculty finds its expression in the large number of languages across our world – the estimates vary between 6,000 and 7,000 depending on how you define a language.Footnote ⁷ At first sight, that might seem a lot of languages, but the peak of linguistic diversity was probably in the fifteenth/sixteenth century, with perhaps as many as two to three times the present number (a rough estimate, given the lack of data). Today, 23 languages account for those spoken by over half the population of the world (close on four billion) and about 40 per cent of present-day languages have a status of ‘endangered language’ (www.ethnologue.com/guides/how-many-languages).

No matter how many speakers a language has, each one is an external manifestation of the internal language faculty which is common to all humans. All humans have the same type of vocal tract and organs of speech. This means that the physical substrate for all the 6,000+ languages on Earth is the same. Not only that, the design features (see Section 20.4 above) are the same for all our languages, again representing a common core of organisation and structure across the Earth.

23.8 Was There One Original Language?

Can one reverse engineer language? Unfortunately not, because even the oldest remnants of language – written documents about 3,000–4,000 years old – already show human language with all the features typical of language today. Older languages are not simpler versions of modern languages.

The idea that all humans once spoke just one language has been around for a long time. Before the dawn of modern linguistics in the early nineteenth century, people were bound to a theological view of the origin of language and thought that Classical Hebrew, the language of the Old Testament of the Bible, was the original language. In modern times, the idea of a single language was resurrected and discussed by the Italian scholar Alfredo Trombetti (1866–1929), who published a book on the subject in 1905.

But was there one original language from which all others derive? That depends on how the evolution of language is viewed. If it is seen as having arisen slowly over a considerable period of time with ever-increasing complexity, there was no one original language, in our modern sense, but rather a series of stages through which humans (Homo sapiens and possibly Neanderthals as well) and their forms of language developed. If, however, you believe the modern language faculty sprang into existence in a super-saturated, language-ready brain (see ‘Grammar: How Did Compositionality Arise?’ in Section 32.5 for more on this view), whatever resulted from this sudden event, in terms of an external manifestation of the language faculty, would have been the original language.

23.9 Language Change

As a research field, language change looms large in linguistics. It would be impossible to do anything near justice to the nuanced insights in this field within a few paragraphs. Here it must suffice to mention a few general principles of change which would likely hold for exolanguages as they move through time.

If we contacted exobeings we would find them at one point in their development. They would have a past behind them, just as we do. The same is true for an exolanguage: it would be a snapshot of its development through time. Assuming the existence of societies in principle comparable to those on Earth, the contact between speech communities from different societies would have led to language mixture over time. This would represent an instance of externally motivated language change, as would that triggered by the behaviour of speakers within a community in their desire to make their speech more like some subgroup, or less so, as the case may be.

There is also internally motivated language change, which is typically found in the transgenerational transmission of language. In this scenario speakers, usually in their early childhood when acquiring their native language, make slight changes to the system on the basis of internal considerations, often to render some set of forms more regular, historically verbs like help : halp > help : helped or noun plurals as in fish : fish > fish : fishes.

The internal and external forces in language change would lead to constant shifts in the form of externalised language across an exoplanet. However, if they had an internal language faculty, as humans have on Earth, this would remain stable and provide the framework within which variation would constantly occur.

On Earth, language diversification probably happened as follows: a language would have arisen in a group by its members using it for communication purposes. Within a community of speakers using this language tiny variations would have appeared over time. These variations would have been passed on to future generations and, given enough time, the variations would have led to major changes, resulting in later generations not necessarily understanding the language of many generations back. The rate of change in natural languages on Earth varies, depending on a variety of factors, such as contact with other groups. However, variation and change is found in all languages. If the variation is minor, we talk about dialects, different but mutually comprehensible forms of one language. Such a dialect becomes a separate language when this mutual comprehensibility no longer holds. An instance of that would be the development of modern Dutch from earlier dialects of German,Footnote ¹⁰ or the diversification of forms of Latin into the Romance languages we know today.

There are other reasons for language change. One very important one is the manner in which children in any given generation construct their language system from what they hear their parents/carers saying around them. If their system is not exactly the same as that which the previous generations internalised in their childhood, change has taken place.

So the genesis of languages is a bit like biological speciation: variation arises and at some later stage interbreeding is no longer possible and an independent species then exists. Similarly, variation in language, for the reasons just outlined, can lead to mutual incomprehensibility and then a new language is taken to exist. Like species, languages can be organised into groups and then larger groups.

The upshot of these considerations is that exoplanets, assuming they have a sizeable population, say hundreds of millions, if not billions, will have many exolanguages as well. Some of these will be spoken by large numbers of speakers and some by smaller groups, just as on Earth, for reasons which have to do with the history of certain countries and the dominance of some groups over others. Exoplanets might also be subject to globalisation which leads to a reduction in variation across societies on a planet. This would in turn lead to most smaller languages disappearing, as is happening rapidly on Earth where the number of languages will be drastically reduced during the course of the twenty-first century.

The ability to speak a language rests on physical aspects of our brains. We can identify areas which are especially important for language, and we can examine individuals with language impairments to gain some insights into the manner in which knowledge of language is stored in the brain. This study of language in relation to the brain is called neurolinguistics. It is a special field which is becoming increasingly a focus of interest for linguists. It is true that it is not possible to pinpoint linguistic activity in the brain, to put the transmission of minute electrical currents between nerve cells in correlation with the production of language. Nor can linguistic structures be assigned to the information stored in these cells. Although the ultimate goal of linking biochemical and electrical processes in the brain with the production and reception of language is still a very distant one, the field is one which has produced significant research results, for instance in the area of aphasia – disturbances in the normal functioning of language, for whatever reason.

24.1 Language Areas in the Brain

The cerebrum is the part of the brain under the skull and is divided into two sections of equal size, called hemispheres (see Section 15.3 above for more details). These are joined by a ‘bridge’ of thick fibres called the corpus callosum. Each hemisphere of the brain controls the opposite half of the body: the left hemisphere is responsible for controlling the right half of the body and vice versa (technically known as contralaterality). The outer layer of both halves is called the cortex (and is greyish in colour when the brain is prepared after death, hence the colloquial term ‘grey matter’ for the brain). The entire brain contains a vast number of neurons (nerve cells) – something in the region of 85–90 billion) – all of which are interlinked by nervous fibres, which allow communication between the nerve cells. The communication away from a nerve cell takes place along axons, which are coated by a sheath of myelin. The pathways to the nerve cell are along dendrites. The central area around the nucleus of the cell is called the soma.

Signals between nerve cells must cross synapses, gaps in the membranes which are found between neurons and axons (Figure 24.1). These synapses are junctions across which information can flow or be impeded, depending on the concentration of chemical substances known as neurotransmitters, which regulate the flow of information through the brain via neurons. These are released when an action potential (an electrical charge) builds up on one side. They cross the synaptic gap and dock into receptors, leading to the action potential being generated on the receiving side. After this there is a re-uptake of neurotransmitters into so-called vesicles, which hold them until required for a future signal. The same principle is used not just in the entire brain, including parts responsible for language, but between nerve cells and muscle cells in various parts of the body.

Figure 24.1 A schematic representation of nerve cells, the inset shows a close-up of a synaptic gap

If neurons are fired for a specific task, such as speech, can one ask if anything is happening in the brain when one is not performing any specific task. The answer would appear to be ‘Yes’. There is a state labelled the default mode network, which basically refers to the state when you are lying there in a wakeful state, not engaged in a specific cognitive task. This state shows blood-oxygen flow in at least the medial prefrontal cortex, the posterior cingulate cortex and the angular gyrus, which is suspended when other regions become active for specific tasks, such as speaking.

Many functions of the brain are associated with one of the hemispheres. The final assignment of these functions to a certain hemisphere is called lateralisation, completed before puberty. This means that a dysfunction of an area of the brain after this watershed cannot normally be compensated for by the transfer of the associated functions to the opposite hemisphere. Puberty is also the cut-off age for acquiring a language with native-like competence.

From investigations of patients with impairments to both speech production and understanding, which were carried out in the second half of the nineteenth century, it is known that there are two key areas in the brain (Broca’s area and Wernicke’s area), along with several others, all associated with language functions (see Figure 24.2).

Figure 24.2 Regions of the human brain (simplified)

24.2 The Binding Problem in Language

Just as there is an issue in binding the various sensory inputs to the brain into a seamless experience of consciousness there a similar problem in both language production and perception in accounting for how pronunciation, grammar and meaning blend in seamlessly when we produce and listen to language. Evidence from the structure of human languages and from language pathology support the ontological status of the different levels of language. They are organised differently but interact closely in the production and reception of language. Just how this happens on the levels of neuronal activity has yet to be demonstrated although it is clear that there are complex circuits of short connections in different cortical regions of the brain and their firing can be tracked in fMRI scans (by tracking blood-oxygen flow). For sounds, the pathways are evident as Broca’s area communicates with the motor cortex/supplementary motor area to initiate pronunciation. But just how meaning and grammar interact and generate sentences on the neuronal level at stages prior to speech initiation is not known (see Section 15.5 above for a more general discussion of this issue).

Imaging studies have revealed that the inferior parietal lobule, a part of the parietal lobe, also called Geschwind’s territory after the American neurologist Norman Geschwind (1926–1984), is strongly connected by nerve-fibre bundles with both the Broca and Wernicke areas. It may thus represent a second, alternative language pathway, parallel to the general pathway between these two areas via the arcuate fasciculus.

24.3 Evidence from Language Impairments

Language pathology is an area of linguistics concerned with language impairments, which can provide insights into the nature and organisation of language in our brains. Pathological aspects of language production and reception are subsumed under the general label aphasia, lit. ‘lack of speech’. There are many subtypes which manifest themselves in different kinds of speech and/or hearing impairment. Insights from aphasia can be used for a number of purposes, for instance, in remedial linguistics where language specialists try to help patients, such as those who have suffered a stroke, to regain as much language ability as possible. But aphasia is also relevant to general linguistics, as the breakdown of language, and how this proceeds, can help us better understand how language in general is structured.

Dementia refers to a decline in cognitive ability due to brain damage or as a result of old age (senile dementia). This usually entails a gradual decline in language performance. Dyslexia is a condition in which individuals often fail to connect the written and the spoken word. Such persons have difficulty with reading and spelling, irrespective of their level of education. The diagnosis of dyslexia is problematic as it occurs to varying degrees and it is notoriously difficult to determine when it is pathological. Dyslexia may be an acquired condition with individuals who were previously literate.

24.4 Types of Aphasia

Aphasiac disorders typically involve damage to either the speech production (Broca’s) or speech reception (Wernicke’s) areas in the brain or perhaps the arcuate fasciculus or the supplementary motor area (see above). When discussing aphasia, scholars use the label fluent aphasia for any kind of language impairment in which the motoric aspect of speech production is not affected (Table 24.1). Where this is the case, the term non-fluent aphasia is used.

Broca’s aphasiacs have typical disturbances in their speech. It is slow and difficult, such speakers seem not to manage grammatical rules, though their vocabulary is normally intact. The type of aphasia where speakers show a lack or confusion of grammatical words by, for instance, omitting small, grammatical words and often inflectional endings is termed agrammaticism.

Speakers with disturbances in Wernicke’s area can speak normally, though what they say often makes little sense. Their sentences are semantically incongruous.

Interestingly, those with Broca’s aphasia are aware that their speech is deficient and tend to respond to remedial treatment whereas Wernicke’s aphasia patients do not, to anything like the same extent, as they do not perceive their speech as deficient. Both types of aphasiacs suffer from the inability to find words, technically known as anomia. Broca’s area lies quite close to that which controls motor movements whereas Wernicke’s area is close to the auditory area. The set of connective fibres between these two areas, the arcuate fasciculus, can also experience damage, causing conduction aphasia, whereby speakers can speak to some extent, retain good comprehension but are incapable of repeating what is said to them.

There is also a rare symptom called isolation aphasia in which the production and comprehension areas become severed from the rest of the brain. Speakers with this impairment can neither produce nor comprehend sentences and only repeat what was said to them or set phrases which were learned in childhood.

It is interesting to note that aphasia affects people’s ability to use sign language just as it does with speech. Different types of aphasia can give us information about the structure of normal language. For instance, people with left-hemisphere strokes may sometimes use nouns but not verbs. This is also true when the nouns and verbs have the same form, so that speakers could understand and produce a sentence like She prefers butter to margarine but not She buttered the toast while it was still hot. This would imply that nouns and verbs are stored differently.

Aphasia can also help answer a key question in current research on language evolution: ‘Is thought possible without language?’ Investigations of persons with global aphasia, such as those reported on by Fedorenko and Varley (2016), with little or no ability to either produce or understand language, can be helpful here. These individuals, despite their near total loss of language, can perform basic arithmetical operations (addition and subtraction), solve problems involving logic, listen to music, engage with the thoughts of others and move around their environments with relative ease. In addition, the investigation showed that healthy individuals when engaging in similar activities do not show firing of neurons in language areas of the brain. This comparison of aphasic with non-aphasic test persons indicates that, for some cognitive activities at least, the language areas of the brain are not active and that thought and language are indeed largely separate. There is a further conclusion to be drawn, this time from the examination of inherited mental disabilities, namely that they do not affect the language competence of the individuals in question to the extent that one might expect. This is additional, if only partial evidence for the separation of cognitive and linguistic abilities.

You have already performed the greatest feat of your life, although you most likely are unaware of it. This is the acquisition of your native language. Within the first few years of your life, you went from nothing to a fully competent speaker of the language(s) you were exposed to. That happens unconsciously, without any instruction,Footnote ¹ in a very short time, with native-speaker competence as the result.

Would the same hold for exobeings? Indeed, would it be valid to assume that exobeings have a division of their lifespans into childhood and adulthood as with humans? Recall that for Darwinian evolution to occur there must be some way for an organism to reproduce and be gradually subject to natural selection. If sexual reproduction applied in the animal world of an exoplanet, reaching sexual maturity would be a feature of animal life. The period before this maturity would be the equivalent to childhood with humans, and the period after maturity is reached would correspond to our adulthood. It is probably fair to assume that the maximum degree of cognitive flexibility would apply before sexual maturity, as this is a developmental learning stage, and hence the acquisition of language would be likely to take place during this key period of an exobeing’s life.

25.1 Are We Predestined for Speech?

It would seem so. Young babies shortly after birth show a tendency to pay attention to speech more than other sounds in their surroundings (this can be shown by tracking eye gaze with infants). It is as if they were predestined to focus on human speech and begin straight away with the job of listening to and dissecting speech as a preliminary to language acquisition.

The acquisition of one’s native language is part of one’s general cognitive maturation. Humans have a relatively long childhood, compared to higher primates, and it has often been speculated that in the evolution of the human species childhood was extended to accommodate the acquisition of a complex language system.

Despite the long human childhood, the acquisition of language is remarkably fast. Childhood lasts until puberty – somewhere between 11 and 13 years of age – but structural knowledge about language is actually acquired in less than half this time. By the age of five or six children have acquired mastery over all the closed classes of their native language, the pronunciation and grammar system. There would appear to be a critical phase in the first year of the child’s life when it becomes attuned to the sounds of the language it is exposed to, with grammar following somewhat later and perhaps extending for the most intricate structures to later years of the first decade of life.

It is extraordinary that language acquisition is a successful process given the poverty of linguistic input and the fact that children receive little or no guidance in language acquisition from those around them. By ‘poverty of linguistic input’ is meant the unstructured and fragmentary language which children are exposed to. Even those children for whom this is true to an extreme extent, like children in homes or orphanages or with absent parents/carers, still acquire their native language as satisfactorily as those who grow up in an intact nuclear family. In contradistinction to second language learning, children are not first presented with simple sentences, progressing gradually to more difficult ones.

Evidence for the innate predisposition to language, the genetic headstart so to speak, can be provided by demonstrating unequivocably that, after acquisition of their first language, individuals have grammatical structures at their command and are aware of mandatory interpretations for which they received no evidence in the linguistic input from their environment during the acquisition process. Consider the following sentences: Fiona believes she is intelligent, Fiona believes her to be intelligent. In the first sentence, she can refer to Fiona or someone else, but in the second it must refer to someone else. Although mothers do not explain such differences to their children, the latter grasp and observe the distinction, given their understanding of underlying grammatical structure.

This brings us to what is called the ‘bootstrapping problem’: all that children are exposed to from birth onwards is a continuous stream of sounds. There are no pauses in this stream, no audible ‘spaces’, to help children segment the phonetic stream. So how do children know what to do? Nobody tells them to organise the sounds they hear around them into words and store these as units of language which they can retrieve later when producing language themselves. And yet this is precisely what children around the world do constantly, no matter what the language of their environment is. The only conclusion one can draw from this is that children have an a priori notion of word; that is, before they hear language, they know what to do with it, how to segment and acquire it. This ability is innate, it is part of the language faculty.

What such observations would also seem to imply is that language acquisition proceeds on the basis of predefined steps, which are determined by innate knowledge about language, this knowledge informing the children about how to manage and structure the information which they pick up about the language in their immediate environment. Another important observation supports this thesis: children correct themselves, given time. For instance, children often start off by treating all verbs as weak and say things like singed, taked, goed, or overapply the s-plural to all nouns, such as foots/feets for feet. Nonetheless, given time, they arrive at the adult forms of their own accord. Such features are technically called ‘errors’ rather than ‘mistakes’, which are irregular and more characteristic of second language acquisition. Errors are non-adult, overgeneralised features which occur because of the stage at which the child is at a given time (acquisition in as yet incomplete). They right themselves with time, when the child appreciates that many word classes contain a degree of irregularity.

The role of caretakers, above all the mother, has led some researchers to claim that their speech – often termed motherese, a deliberately simplified form of the adult language – is important for the child’s acquisition of language. Opinions are still very much divided on this point and the evidence of neglected children who acquire language normally would point away from any importance of motherese, despite its obvious value as evidence of the mother’s affection and concern for the child.

25.2 The Absence of Exposure to Language

If children are not exposed to language during the first decade of childhood, they will not acquire it later. There have been some instances where this has happened, most of them also involving extreme traumatisation of the children. A well-known case from the mid-twentieth century is that of a girl, just called Genie to protect her true identity. She was kept confined to a single room, physically abused by her psychotic father and exposed to no language whatever until the age of 13, when she was seen by chance when her mother went with her to a local council building, stumbled into the wrong room and some staff immediately noticed the girl’s extremely neglected and retarded state. After she was liberated from her domestic situation Genie was accompanied by psychiatrists and linguists in an attempt to provide a new world and a semblance of normality for her. She made some initial progress in general behaviour but none of any substance in language acquisition. Whatever the role her decade-long abuse played in the stunting of her personality, it became obvious that Genie would not acquire language to any degree comparable with other children. While she did manage to learn some words, she had no command of grammar.

An older case is that of Victor of Averyon (c. 1788–1828), a young boy who was found in 1800 in woods in southern France and assumed to be a feral or ‘wild’ child, who spent his life hitherto surviving on his own in the countryside without any human contact. Victor’s case was studied in some detail by the French physician Jean Marc Gaspard Itard (1774–1838), who had taken Victor into his home and accompanied him for several years. Again, the progress made in language acquisition was truly minimal. Victor could only speak a few words and could not create any sentences or use French grammar in any way. There has been much discussion of this and other supposed cases of feral children. Nowadays many scientists doubt if such children survived in the wild since infancy, as that would have been well-nigh impossible. In the case of Victor, it is assumed that he was abused by his parents before running away, at which time he must have been at least six or seven, otherwise he could not have survived on his own in the wild.

The details and discussions surroundings such cases, and others like them, are not relevant to the present treatment of language. However, what they do show is that if children do not have exposure to language in the first decade of their lives then the window of opportunity closes and after that they cannot master grammar but can learn some words. This supports the Critical Period Hypothesis in linguistics (actually a hypothesis about a group of sensitivity periods), which maintains that language acquisition must take place and be completed before puberty if normal levels of language competence are to be attained.

The American linguist Susan Curtiss, who accompanied Genie for several years, is convinced that she had cognitive abilities on a par with normal intelligence but without a mastery of language. This ascertainment is of relevance to the ongoing discussion in linguistics of the relationship of language to thought. If children who have been deprived of the opportunity of normal acquisition in childhood nonetheless show approximately normal cognitive competence, this would support a weak connection between language and thought, if not indeed go a good way to confirming that the latter is possible without the former.Footnote ²

25.3 Characteristics of Language Acquisition

The ability to learn one’s native language is part of one’s genetic endowment, ultimately encoded in DNA.Footnote ³ We have an instinct to acquire language, which is activated by birth and unfolds in infancy just as does the instinct to walk upright. Just like walking, it takes a while to get going but it will continue automatically, circumstances allowing. Thus, language acquisition can be compared to other instincts such as that to use one’s hands or to develop telescopic vision or binaural hearing.

Intelligence has no direct bearing on acquisition, and children of different degrees of intelligence all go through the same process of acquiring their native language, although individuals can and do differ in their ability to manipulate open classes, such as vocabulary, and in the style of written language they master consciously.

25.4 Stages of Language Acquisition

Language acquisition consists of a sensorimotor component, responsible for pronunciation (phonetics/phonology), and a cognitive component, responsible for the non-physical aspects of language, the grammar and vocabulary. Children pass through clear stages of acquisition in the first five or six years of their lives. Within each of these stages there are recognisable characteristics. For instance, up to the two-word stage, nouns and/or verbs mainly occur. No children begin by using conjunctions or prepositions, although they will have heard these word classes in their environment. Another characteristic is overextension. Children always begin overextending, for example in the realm of semantics, by using the word dog for all animals if the first animal they are confronted with is a dog. Or by calling all males daddy or by using spoon for all items of cutlery. The conclusion one can draw from this behaviour is that children move from the general to the particular. To begin with their language production is undifferentiated on all linguistic levels (Table 25.1). With time they introduce more and more distinctions as they are repeatedly confronted with these from their surroundings. Increasing distinctions in language may well be linked to increasing cognitive development: the more discriminating the children’s perception and understanding of the world is, the more they reflect this in language.

These divisions of the early period of first language acquisition are approximate and vary among individuals. It has been noted that bilingual children start speaking slightly later than their monolingual counterparts, most likely because they have to absorb information from two languages before beginning to reproduce these.

25.5 Abduction and Ambiguity in Language

In progressing through the stages of language acquisition children use a form of imprecise logic called abduction in which they choose the easiest and most likely explanation. Consider the following instance of abduction, consisting of two premises, but a conclusion which may well be false: (i) tigers have stripes, (ii) this animal has stripes, therefore: (iii) this animal is a tiger. Not necessarily, there are animals with stripes that are not tigers, such as zebras. Now consider this example from language: (i) adverbs end in -ly, (ii) this word ends in -ly, therefore (iii) this word is an adverb. Not necessarily, there are adjectives which end in -ly, such as friendly, likely, as in a friendly letter or a likely story. While children use abductive logic as a working hypothesis during acquisition, they revise their assumptions as they are exposed to evidence which contradicts these.

Languages contain ambiguity, some quite a lot if they do not formally mark grammatical categories or agreement patterns. However, speakers can handle this ambiguity because context and expectations regarding likely interpretations assist them. Research by the American linguist Barbara Lust (Cornell University) has shown that children under the age of three can already deal with a grammatical form which has more than one semantic interpretation, such as Ernie took a bite of his banana and Bert did too, where the preferred assumption by the test children is that each took a bite of their own banana. This interpretation is probably prompted by experience of the world: usually people eat their own and not each other’s food. Our cumulative experience during childhood results in a store of encyclopedic knowledge about the world, which later guides us in making correct interpretations of vague or ambiguous language. Significantly, artificial intelligence does not have a comparable store of knowledge or direct experience of the world; this is the main obstacle to constructing software which can provide semantically acceptable interpretations of linguistic input.

25.6 Localisation of Language and Early Childhood

At the very beginning of language acquisition, in the first year and a half, neuronal activity when producing words and reacting to speech tends to be spread across both hemispheres of the brain. With the increasing knowledge of language, the left hemisphere becomes dominant, with the right hemisphere finally being removed from the process of language perception and production. However, this fixation on the left hemisphere is not unalterable. There are reported cases of children who had their left hemisphere removed or disconnected, usually to put an end to debilitating seizures, with the loss of language. However, within a couple of years such children can show progress is utilising the right hemisphere for the acquisition of language.

25.7 Language Transmission

Language is obviously passed on from parents to their children. But on closer inspection one notices that it is the performance (language in use) of the previous generation which is used as the basis for the competence (abstract knowledge of language) of the next. To put it simply, children do not have access to the linguistic knowledge of their parents. An initial, simplified form of this transmission can be formulated.

Language competence is the abstract ability to speak a language, meaning that knowledge of a language is independent of its use. It is constructed during early childhood by combining innate knowledge of language in general (in the language faculty) with actual linguistic input from the child’s surroundings. Performance is the actual use of language. Its features do not necessarily reflect the speaker’s language competence. For example, when one is nervous or tired one may have difficulties speaking coherently, but one’s competence remains.

The three stages in transmission listed above account for why children can later produce sentences which they have never heard before: they commit the sentence structures of their native language to long-term memory and have a store of words as well. When producing new sentences, they use the internalised structures and the vocabulary they have built up for themselves. This process allows children to produce a theoretically unlimited number of sentences in their later lives. However, certain shifts may occur if children make incorrect conclusions about the structure of the language on the basis of what they hear. Then there is a discrepancy between the competence of their parents and that which they construct. This is an important source of language change.

25.8 The Logical Problem of Acquisition

The following problem exists with regard to early language acquisition: it would seem impossible to learn anything about a certain language without first already knowing something about language in general. That is, children must know what to expect in language before they can actually order the data they are presented with in their surroundings, extract structures from sentences and ascribe meanings to the words they encounter.

25.9 The Evidence of Pidgins and Creoles

Children who have very poor input in their surroundings tend to be creative in their use of language. Grammatical categories which they unconsciously regard as necessary but which are not present in the input from their environment are then created by the children. This has happened historically in those colonies of European powers where a generation was cut off from its natural linguistic background and only supplied with very poor unstructured English, Spanish, Dutch, etc. as input in childhood. Such input, known technically as a pidgin, was then expanded and restructured grammatically by the children of later generations and is known in linguistics as a creole. Here one can see that, if the linguistic input from their environment is deficient, children create the structures which they sense are lacking, going on their own innate knowledge of language.

The implication of the above situation is that children look for language and if they do not find it, they create it somehow, so that they can externalise their internal linguistic knowledge. In this sense language is a true instinct because it starts to unfold of its own accord and does not need to be consciously triggered. During the process of creolisation children add considerable regularity to the largely unstructured pidgin input of previous generations (see ‘Iterated Learning’ in Section 32.5 below).

25.10 Is There a Gene for Language?

The simple answer to that question is ‘No’. Much has been made in the media of the discovery of the FOXP2 gene (see Section 24.4 above). This is responsible for the acquisition and control of vocalisation and is closely connected to our ability to speak. However, it is not the sole gene implicated in language. In fact, it is only to be expected that a system as complex as human language would not be triggered by the transcription/translation of a single gene.

The FOXP2 gene was also present in Neanderthals, though not in identical form. It would, however, imply that they had a comparable degree of vocal control, which would have provided them with one of the preconditions for language.

25.11 Constructed Languages

In the late-nineteenth and at the beginning of the twentieth century various scholars turned to the task of devising an artificial language, which could be used as a general means of communication among speakers who do not understand each other’s language. This function is fulfilled by English nowadays but before the spread of the latter as a worldwide common language (technically called a lingua franca), other practical suggestions were put forward about how to fill the perceived gap.

A proposal which enjoyed brief, if intense, popularity was Volapük ‘world talk’ (in this language), put forward by the German priest Johannes Martin Schleyer in 1879. He devised the language ‘out of pure love for troubled and divided humanity’. It was initially very successful, with conferences and public discussions. Even an academy for Volapük was founded in Paris in 1889. However, the language was very cumbersome with too many endings. The community of Volapük supporters found the language too much of an effort to learn and, around 1890, most of them switched allegiances to the following.

Esperanto is another artificial language invented by the Polish scholar Ludwig Zamenhof (1859–1917), again in the late-nineteenth century. It was intended as an easy-to-learn, regular language, which would, like Volapük, further international communication and understanding. The language is based on Romance elements and was intended to be easy for Europeans to learn. The name clearly suggests Spanish esperanza ‘hope’. Of all the various proposals for an artificial language, Esperanto is the only one which can be said to survive to this day. Although it is not used widely, there is nonetheless a dedicated community of Esperanto users worldwide.

In the course of the twentieth century, various modifications of Esperanto have been put forward, largely to rid it of what were perceived as unnecessary difficulties in the language. Even the great historian of English, the Dane Otto Jespersen (1860–1943), offered a greatly modified version called Novial, which, like so many others, did not catch on.

Various suggestions have also come from Romance scholars, such as Interlingua, to devise a sort of common-core language which would be comprehensible for all speakers of Romance languages.

In the English-speaking world, attempts have been made to provide non-natives with simplified versions of the language for rudimentary communication. The English literary scholar C. K. Ogden put forward his proposal for Basic English in 1930. This consisted of only 850 words, but interest in the suggestion quickly waned because it was impracticable.

And then there are languages created for science fiction. The best known of these is Klingon, used by the Star Trek franchise as the language of a group of particularly belligerent people within the series of films. It was devised by the American linguist Marc Okrand in the 1980s and expanded in later years, leading to a considerable community of enthusiasts for the language. Klingon only uses sounds found in human languages but in combinations which are not indicative of any single language and which contribute to its overall strangeness. It is characterised by many retroflex and uvular sounds giving it a general ‘throaty’ character and has a syntax which contrasts with English, for example, by using the basic word order object–verb–subject (O–V–S). This is very unusual cross-linguistically, attested for the indigenous South American Carib language Hixkaryana, spoken by only about 500 people around a tributary to the Amazon river. However, it is common as a highlighting word order in many languages, for instance German, as in Dieses Auto möchte er, nicht das andere, literally ‘this car (O) would like (V) he (S), not the other one’. Marc Okrand has said that he deliberately chose the word order OVS for its unusualness among human languages. Such a bias is rooted in the fact that the constructed language Klingon was tailored to the requirements of the entertainment industry and was not in any way intended by its author as an attempt to represent how he thought actual exobeings on a real exoplanet might speak, should they exist.

On our Earth the dividing line between animals and humans is determined by the ability to speak. Assuming that evolution is the only manner by which complex biological structures can arise in our universe, the exobeings of any exoplanet must also have arisen through an essentially similar process of evolution as we humans have.

Language separates us from the rest of the animal worldFootnote ¹ and makes us what we are, beings who live in intricate social networks, with considerable cognitive powers and the ability to convey knowledge from one generation to the next and thus build up vast bodies of knowledge which extend far beyond what a single individual could achieve in a lifetime. All this is possible because we can formulate our thoughts in language which others can then understand and engage in exchanges. Without this foundation there would be no science and none of the technological advances which we enjoy in today’s world, built on the discoveries and insights of science. And language is used for much more, for the expression of literature, for the recording of history and for the myriad types of communication in which humans are involved in societies across the globe.

The basic question here is whether animal communication, specifically that found with higher primates, and communication among humans is different in kind or just in degree. Is there a continuum between our systems of communication and theirs? The research community is divided on this issue.

26.1 How Intelligent Are Animals?

Intelligence is difficult to define satisfactorily, let alone measure in the animal kingdom. There are different kinds of intelligence, the main two types being general and emotional. General intelligence among animals can be seen in their ability to perform cognitively challenging tasks and/or to influence their environment to their advantage. A well-known example is that of New Caledonian crows, which are able to prepare a stick with a hook-like end to extract grubs from places not easily accessible. They also show this capacity in captivity and display a learning curve in mastering such tasks.

Signs of emotional intelligence can be observed in advanced mammals, two of which are empathy and mourning. Empathy is feeling with other beings, subjectively appreciating their predicaments. Dolphins show empathy across the species barrier and have been known to assist humans who get into difficulty when out of their depth in water. Many dogs also show empathy, especially intelligent ones like Newfoundlanders, who can rescue humans from drowning. It could be objected that such dogs see all humans as potential food providers and hence have an intrinsic motivation to save them, but for dolphins (living in the wild) no such assumption can be made.

Humans mourn their dead. Cultures have rituals they engage in when a member of their group dies. Burial rituals can take on a variety of forms around the world and are usually connected with taking leave of dead individuals before they depart for ‘the other world’. Such leave-taking involves mourning, spending time around the body of a deceased person and experiencing sadness at their death. Such behaviour has been observed with dolphins and many advanced mammals, such as elephants and chimpanzees, especially on the death of one of their young. Many non-mammals, such as egg-laying tortoises or spawn-laying frogs, do not stay around for their many young to mature and so would not even know if they were alive or not.

Mourning the dead is connected with aging in populations. If the latter contain older members, death by natural causes is common and mourning, possibly with rituals, will also be widespread. What role age plays in animal groups can help us determine their probable cognitive powers. Some groups of mammals may consist of more than two generations, such as elephants, certain whales and monkeys. In these cases older females (‘grandmothers’) can provide care for their grandchildren and the group can further benefit from the experiences of the older animals, for instance in procuring food. Transgenerational animal behaviour requires bonding beyond a single animal’s own parents and implies higher levels of cognition, an issue which has been studied in detail by Anne Innis Dagg (see Reference Innis DaggInnis Dagg 2009).