Undecidability of English Syntax and Semantic Feedback

There are three main components to any form of language communication: syntactic, semantic and pragmatic. Due to the immense impact that syntax has upon semantics, linguists have always been interested in the relationship between the first two. Although agreement exists regarding many points in connection with this, others, maybe just as important, remain open to debate.

The problem of syntactic ambiguity and “undecidability” in English is nothing new. However, starting from the idea that there is a two-way relationship between the syntactic and semantic component of language production, a theory of how syntax cannot be determined in the absence of semantics has become very popular. Practical implications of this include computer programmer Jeffrey Kegler’s demonstration, which will be analysed below, that there exist sentences which are unsyntactic “if and only if they are syntactic”.

My opinion is that such a theory and its implications should not be accepted without question, due to the fact that there are issues to be identified with it. Since this is difficult to disprove on a universal scale, in the present article I have taken up as case study Kegler’s demonstration and tried to point out its flaws.

The syntax-semantics link is very clearly seen even from the preliminary considerations that by the syntactic component of language production we mean how morphemes work together in the forming of sentences as well as their roles in them, word order etc. and by the semantic component we refer to different aspects of the meaning(s) sentences have. Although some theories presume the contrary, this article is written in the view that the syntactic component ‘happens’ first and the information it offers is used by the semantic one to create meaning.

A very important aspect must be established before analysing anything else related to the impact of semantic feedback on the undecidability of syntax.  Let us take a look at the following two examples:

  1. Green dreams sleep furiously.
  2. Airplane from to the and for of or pilot up.

Neither of the above sentences has any meaning. However, it is important to notice that, while the second one has no meaning in the absence of syntax, the first one also has no meaning but in the presence of syntax. What we conclude is that we can have syntax without meaning, but we cannot have meaning without syntax. This is unquestionable.

Jeffrey Kegler considers that we can build sentences that are unsyntactic if and only if they are syntactic based on semantic feedback (i.e. the ability of semantics to impact upon a sentence’s syntax by solving ambiguities).  Let us take a look upon his demonstration and analyse it.

He starts from the morpheme “closed”, which can either be a past participle or an adjective. Thus, a sentence such as “The window was closed” is syntactically ambiguous since we cannot decide if the morpheme “closed” is a subject complement or part of the passive voice. To solve this, Kegler expands his example into “The window was closed and the door opened by the same person”. In the latter example, it is clear that the morpheme “closed” is a passive and the sentence is no longer ambiguous because of the meaning of the added elements. We have an agent which opened the door and closed the window. Semantics solved the ambiguity.

Afterwards, he gives an example, noted (U): “The window was closed and the door opened by the burglar after he discovered that the window was in fact a beautifully executed trompe d’oeil mural”. He explains that (U) is syntactically correct. However the window is fake, meaning that the burglar couldn’t have closed it. Thus, “closed” cannot be a passive even if we have “and the door opened by the burglar” following it in the word order. Therefore, he says, after the semantic feedback, (U) is no longer syntactic. As such, (U) is unsyntactic if it has semantics and since semantics cannot exist without syntax, (U) is unsyntactic if it has syntax.

The first issue with this demonstration is that (U) is unsyntactic if it has semantics comes into contradiction with the fact that we cannot have meaning without syntax. If meaning is involved, we may never claim that the structure whose meaning we are talking about has no syntax. My personal opinion is that (U) has no meaning and it’s syntax was ambiguous, or at least bad, from the start. We may also consider that its syntax is good but the sentence is unable to provide proper meaning. Either way, semantic feedback cannot make this sentence’s syntax impossible to decide because its overall meaning is off.

Furthermore, let us look back on the example “The window was closed and the door opened by the same person”. It is true that, unlike just “The window was closed” , this one is no longer syntactically ambiguous. The ambiguity was resolved by coordination and introduction of the agent, as Jeffrey Kegler himself admits. It is not the semantic component that had an impact here. Thus, the careful linguist cannot conclude, not from this example anyway, that syntax can be decided by semantics. This is also proof that the second part of the demonstration presented above is wrong.

Embracing the theory that syntax cannot be decided in the absence semantics, Jeffrey Kegler tries to prove that there are sentences which lack syntax if and only if they have syntax, or, to rephrase that, that semantic feedback may render impossible a decision on the syntactic/unsyntactic character of a sentence. As shown, strong reasons to believe that is wrong exist. The programmer’s demonstration along with my opinion on it perfectly reflects how easily matters can fall to interpretation. Although, I do not say that semantics may never have an impact upon syntax, I think the theory that syntax can never be determined without semantics should not be accepted as true. Consequently, I do not believe that meaning may, in some cases, make it impossible to decide whether a sentence is syntactic or not.

Further Reading

Kegler, Jeffrey. “The Syntax of English is Undecidable”.  Ocean of Awareness. Movable Type. Web. 15 may 2014. http://blogs.perl.org/users/jeffrey_kegler/2012/03/the-syntax-of-english-is-undecidable.html

Preda Alina, Trifan Manuela, On Text Production: the integral relationship to translation, PRESA UNIVERSITARǍ CLUJEANǍ, CLUJ-NAPOCA, 2008


Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s