The syntactic structure of sentences is known to be one of the many factors responsible for complexity of prose text. In the following well-known experimental findings by psycholinguists, it was observed that subjects found certain structures to be more complex than others:
NOTE: '>' indicates greater complexity, and deleted phrases in square brackets indicate Gaps
- Doubly self-embedded clauses > Singly self-embedded ([1], [2])
- Eg. 'The pen the author the editor liked
[the author] used [the pen] was new.' ([3]) vs. 'The pen the author used [the pen] was new.'
- Reduced-form Relative clause > Non-reduced form ([3])
- Eg. 'The pen the author the editor liked
[the author] used [the pen] was new.' vs. 'The pen that the author whom the editor liked [the author] used [the pen] was new.' ([3])
- Complementizer 'that' deleted in CP complement > 'that' present ([4], [5])
- Eg. 'They believed the earth was flat.' vs. 'They believed that the earth was flat.'
- Object-extracted gap > Subject-extracted gap (various - see, for e.g., [6])
- Eg. 'The hunter the wolf bit
[the hunter] was taken to the hospital.' vs. 'The wolf that [the wolf] bit the hunter was shot in the forest.'
- Reduced-form Passive > Unreduced-form Passive > Active ([7])
- Eg. 'The king tutored
[the king] by Aristotle was Alexander.' vs. 'The king who was tutored [the king] by Aristotle was Alexander.'
- Sentential subject > Sentential complement ([8], [9])
- Eg. 'That the picture that he painted
[the picture] was pretty annoyed her.' vs. 'It annoyed her that the picture that he painted [the picture] was pretty.'
- Sentential subject in Relative > Relative in sentential subject ([10], [11])
- Eg. 'The king whom the fact that his army caused great suffering saddened
[the king] was Ashoka.' vs. 'The fact that the army that he sent [the army] caused great suffering saddened Ashoka.'
- Passive > Active ([12])
- Eg. 'The king who was tutored
[the king] by Aristotle was Alexander.' vs. 'Aristotle tutored Alexander who [Alexander] was a king.'
The term 'garden path' sentence refers to sentences in which the initial words of the sentence lead the reader to certain expectations that are later belied. This causes some 'confusion' as the sentence needs 'reanalysis', sometimes from the beginning of the sentence. 'Garden path' sentences can be syntactic or semantic in nature. Syntactic 'garden path' sentences usually involve missing 'structural cues'; these can be identified by a sophisticated Deep Parser. Semantic 'garden path' sentences, on the other hand, are difficult for a Deep Parser to identify because these sentences are usually 'well-formed'; hence this problem is best left to some other 'reasoning-about-the-world' AI program.
In 'real world' sentences, a combination of the above complex structures can be found in a single clause/ sentence. However, very little experimental evidence is available concerning the perceived complexity of such combinations of complex structures (see for e.g., [13]). This is a problem for a method that rates the perceived complexity of 'real world' sentences. However, such a method adds value by identifying all the complex syntactic structures that exist in the sentence; it is then up to the author to take appropriate corrective measures.
References
- [MillerChomsky1963] Miller G., Chomsky N.. Finitary models of language users. Handbook of Mathematical Psychology. 1963;2:419-491.
- [ChomskyMiller1963] Chomsky N., Miller G.. Introduction to the formal analysis of languages. Handbook of Mathematical Psychology. 1963;2:269-321.
- [FodorGarrett1967] Fodor J., Garrett M.. Some syntactic determinants of sentential complexity. Perception & Psychophysics. 1967;2.
- [Hakes1972] Hakes D.. Effects of reducing complement contructions on sentence comprehension. Journal of Verbal Learning and Verbal Behavior. 1972;11:278-286.
- [OsterhoutEtAl1994] Osterhout L., Holcomb P., Swinney D.. Brain potentials elicited by garden-path sentences: Evidence of the application of verb information during parsing. Journal of Experimental Psychology: Learning, Memory and Cognition. 1994;20:786-803.
- [GordonHendrickLevine2002] Gordon P., Hendrick R., Levine W.. Memory load interference in syntactic processing. Psychological Science. 2002;13:425-430.
- [RaynerCarlsonFrazier1983] Rayner K, Carlson M, Frazier L.. The Interaction of Syntax and Semantics during Sentence Processing: Eye Movements in the Analysis of Semantically Biased Sentences. Journal of Verbal Learning and Verbal Behavior. 1983;22:358-374.
- [FodorBeverGarrett1974] Fodor J., Bever T., Garrett M.. The Psychology of Language. New York: McGraw-Hill; 1974.
- [TaborEtAl1997] Tabor W., Juliano C., Tanenhaus M.. Parsing in a Dynamical System: An Attractor-based Account of the Interaction of Lexical and Structural Constraints in Sentence Processing. Language and Cognitive Processes. 1997;12:211-271.
- [Gibson1991] Gibson E.. A Computational Theory of Human Linguistic Processing:Memory Limitations and Processing Breakdown. Carnegie-Mellon University; 1991.
- [Gibson1998] Gibson E.. Linguistic complexity: locality of syntactic dependencies. Cognition. 1998;68:1-76.
- [ForsterOlbrei1973] Forster K., Olbrei I.. Semantic heuristics and syntactic analysis. Cognition. 1973;2:319-348.
- [Gouvea2000] Gouvea A.. Working Memory and Theories of Syntactic Complexity.. 2000.