Linguistic Theories about Agrammatism. The Trace Deletion Hypothesis and the Tree Pruning Hypothesis

Essay, 2015
7 Pages, Grade: 2,0


1. Introduction

Agrammatism is "the name given at the beginning of the [nineteenth] century to a pattern of language production that appeared to lack grammatical structure."[1] It has been defined as a “tendency to form sentences without the correct inflectional structure”[2]. This disorder is a result of Broca's aphasia[3], which is caused by damage to Broca's area. “Agrammatic patients speak effortfully using a telegraphic style, often omitting (or substituting) grammatical morphemes and function words, while still using content words rather appropriately.”[4]

However, agrammatism is not limited to the production of language; it is “a reduced ability to understand or produce most grammatical markers”[5]. Therefore, it is a disorder which affects both production and comprehension. “While there is good performance on comprehension of simple actives and subject gap relative clauses, agrammatics show poor performance on reversible passives and object-gap relative clauses."[6]

In the following essay, I will present one linguistic account that concerns itself with the comprehension deficit and one theory which focuses on the impaired production connected to agrammatism.

2. The Trace Deletion Hypothesis by Grodzinsky

The Trace Deletion Hypothesis (TDH) was first proposed by Grodzinsky in 1984 and was severely restricted in 1995.[7] The following chapter intends to describe both the original version of the TDH and the restrictions made.

The original TDH states that “all traces of movement are deleted from [Surface]-structure representation”[8], defining agrammatism as a “[p]artial syntactic deficit”[9]. In the syntactic theory underlying this account, traces are usually present as a connection between a moved constituent and their location in the deeper structure. If those traces are deleted, the theta roles cannot be assigned correctly.[10] This account states that the only syntactic impairment is the deletion of traces; everything else in the structure is intact[11]. If only traces are impaired in agrammatism, then only the process of “projecting lexical material onto sentence structure and linking Θ-roles to NPs […] and nothing else […] is disrupted”[12]. Since agrammatism is a feature found in Broca's aphasia and Broca's aphasia is caused by the impairment of Broca's area, it follows that Broca's area`s sole function must be the production of traces. If Broca's area had any other function concerning the computation of syntax, there would be further impairment visible in agrammatism. This account, therefore, assigns a fairly simplistic function to Broca's area.

Grodzinsky himself acknowledges the limitations of the TDH by stating that the deletion of traces in itself does not explain why an agrammatic patient is unable to retrieve the correct theta role from the amount of unimpaired information still available to him or her.[13] In order to explain theta role assignment errors in specific structures, that is chance performance in comprehending passives and object gap relatives, for example, Grodzinsky added a non-linguistic, cognitive strategy to his account, namely the Default Strategy.[14] This strategy assigns a default theta role to moved NPs with deleted traces. In the tests conducted, this theta role was always the agent role.[15] In the case of passives and object gap relatives, among others, this strategy leads to two competing agent theta roles, one assigned through the cognitive strategy and the other assigned grammatically, for example through the function word “by” in a passive.[16] Since only the information concerning traces, and no other information, is deleted, the agrammatist is aware of the fact that the agent theta role can only be assigned once in a clause. In consequence, the agrammatic patient has to guess which assignment of the agent theta role is the correct one, and this causes the chance performance in comprehension.

On the other hand, agrammatists perform above chance in the comprehension of structures like subject gap relatives. This is explained by the fact that the moved NP in this case is supposed to receive the agent theta role. Therefore, even though the traces are deleted as well, “the strategy correctly compensates by assigning that NP the Agent role by default.”[17]

As Grodzinsky states, this theory is “a sweeping generalization[, claiming] that any trace, in any position, is deleted”[18]. New empirical findings at the time and changes concerning the theory of syntax[19] led to a restriction of the original account in 1995, “making it more precise”[20]. This revised TDH consists of two restrictions. The first is that trace deletion is limited to theta positions, the second states that the applied cognitive strategy only applies to referential NPs[21], “referentiality being defined relative to a discourse”[22]. Thus, the TDH was restated in 1995 as: “only traces in Θ-positions are deleted, and NPs lacking a Θ-role receive one strategically iff they are referential.”[23] The first restriction was included due to findings suggesting that “verb movement […] appears to be preserved in agrammatism”[24]. The second restriction is based on two asymmetries. The first concerns a change in comprehension performance as soon as a psychological verb is the predicate of a passive construction; here, the comprehension level changes from chance to below chance.[25] The second asymmetry occurs in the discrepancy of comprehension levels in confrontation with Which and Who questions. While Which questions lead to a chance performance, in line with the original TDH, Who questions result in a seemingly normal performance.[26] The observation that Which requires “reference to previous discourse”[27] and therefore “both syntactic and contextual information [, while Who questions] can be asked without presupposition”[28] leads to the mentioned second restriction involving referentiality. In the paper cited, the Default Strategy is thus referred to as R-strategy, or Referential-Strategy, because it excludes non-referential items[29].


[1] Levy 105

[2] (20.12.15)

[3] See (20.12.15)

[4] Levy 105

[5] (20.12.15)

[6] Levy 105

[7] See Grodzinsky, 27

[8] Grodzinsky 33

[9] Grodzinsky 29

[10] Grodzinsky 33

[11] Grodzinsky 33

[12] Grodzinsky 48

[13] Grodzinsky 33

[14] Grodzinsky 33

[15] Grodzinsky 33

[16] Grodzinsky 34

[17] Grodzinsky 34

[18] Grodzinsky 36

[19] Grodzinsky 27

[20] Grodzinsky 27

[21] See Grodzinsky 28

[22] Grodzinsky 47

[23] Grodzinsky 28

[24] Grodzinsky 28

[25] See Grodzinsky 28

[26] See Grodzinsky 28

[27] Grodzinsky 41

[28] Grodzinsky 42

[29] See Grodzinsky 44

Excerpt out of 7 pages


Linguistic Theories about Agrammatism. The Trace Deletion Hypothesis and the Tree Pruning Hypothesis
University College Dublin
Seminar: Language Impairment
Catalog Number
ISBN (eBook)
Aggramatism, Language Impairment, Linguistics, Language Production, Language Comprehension, Broca's aphasia, Broca's area, Grodzinsky, Trace Deletion Hypothesis, Friedmann, Tree Pruning Hypothesis, TDH, TPH
Quote paper
Silvia Schilling (Author), 2015, Linguistic Theories about Agrammatism. The Trace Deletion Hypothesis and the Tree Pruning Hypothesis, Munich, GRIN Verlag,


  • No comments yet.
Read the ebook
Title: Linguistic Theories about Agrammatism. The Trace Deletion Hypothesis and the Tree Pruning Hypothesis

Upload papers

Your term paper / thesis:

- Publication as eBook and book
- High royalties for the sales
- Completely free - with ISBN
- It only takes five minutes
- Every paper finds readers

Publish now - it's free