At first glance, the question in the title might seem like a paradox. If the answer is ‘yes’, and impossible languages do exist, then those languages aren’t impossible at all. However, the title aims to address a different question: is there a limit to the variety of languages that there could be? In other words, can we think of a language that will never be spoken by humans?
In order to know what an ‘impossible language’ is, we first need to know what a ‘possible language’ is. If we look at all the different languages around the world (current estimates are around 6000 to 7000), there seems to be unlimited variation in the number of possible languages. Indeed, for a long time there have been heated debates about the question whether there are universal properties that are shared by all languages. To evade this discussion, we can also reverse the question: are there universal properties of which we know that none of the languages have them? If the answer is ‘yes’, then we have found a possible property of an impossible language.
Let us picture a situation whereby an alien has landed in England, and he tries to figure out what the English people are constantly saying to each other. The alien has access to a dictionary, which allows him to translate all words, but he doesn’t yet know the grammatical rules. After being in England for some time, the alien hears the combination of a normal sentence and a yes-no question:
- The boy is very happy.
- Is the boy _ very happy?
The alien notices that the question in 2 was made by placing the word ‘is’ at the start of the sentence (the ‘_’ in sentence 2 indicates the position of ‘is’ in sentence 1). The alien formulates the following grammatical rule: in order to make questions in English, you have to place the third word at the start of the sentence. Let’s call this rule the ‘third-word-rule’. Next, the alien again hears a combination of a normal sentence and a question:
- The tall boy is very happy.
- Is the tall boy _ very happy?
In the question in 4, the word ‘is’ has again been placed at the start of the sentence. This is the fourth word in sentence 3. The third-word-rule appears to be wrong, because this rule would have placed the word ‘boy’ (the third word in sentence 3) at the start of the sentence, and would thus have created a very weird sentence:
- Boy the tall _ is very happy?
The alien has already heard two questions and tries to construct a new grammatical rule that can explain both examples. He formulates a first-verb-rule, which says that questions are made by placing the first verb at the start of the sentence. But then the alien hears the following sentence:
- The tall boy, who is a student, is happy.
Using the first-verb-rule, he tries to create a question by placing the first verb at the start of the sentence, which yields:
- Is the tall boy, who _ a student, is happy?
This sentence is clearly ungrammatical, and people give the alien a strange look. After a while, he realizes what the correct rule is: in order to make a question in English, you place the verb that comes after the subject at the start of the sentence. This rule does not refer to individual words, but rather to groups of words, also called ‘constituents’. The subject of the sentence is such a constituent. The grammatical rule doesn’t care about the number of words that constitute the subject: ‘the boy’, ‘the tall boy’, or ‘the tall boy, who is a student’. This example illustrates that you don’t make questions in English by counting words. This doesn’t only apply to English, however, or even to this specific construction. There are no languages around the world that have counting rules, which make sentences by referring to the linear position of a word in the sentence.
Counting rules would thus be a possible property of an impossible language. It would even be a plausible property of a language. However, several studies have shown that people find impossible languages (that is, languages with impossible rules) unnatural. To give an example, brain research led by the linguist Andrea Moro has shown that, when participants get better at possible languages (languages with possible rules), brain activity in important language areas increases. Interestingly, activity in those same areas decreases when participants get better at impossible languages (such as those with counting rules).
There are no languages around the world that have counting rules, which make sentences by referring to the linear position of a word in the sentence.
The foregoing discussion shows that impossible languages indeed exist. It also shows that the properties of possible languages are not simply a collection of rules that are shaped by our cultures, but rather that they are part of our biology. A linguist who creates an impossible language can thus be compared to a biologist who creates an impossible animal (or an impossible human). This approach to language research can give us important insights into what makes us human.
Want to know more about impossible languages?
Moro, Andrea (2015). Boundaries of Babel. The brain and the enigma of impossible languages. MIT Press.
Moro, Andrea (2016). Impossible languages. MIT Press.