Lecturer's Précis - McCarthy and Warrington (1984)

Copyright Notice: This material was written and published in Wales by Derek J. Smith (Chartered Engineer). It forms part of a multifile e-learning resource, and subject only to acknowledging Derek J. Smith's rights under international copyright law to be identified as author may be freely downloaded and printed off in single complete copies solely for the purposes of private study and/or review. Commercial exploitation rights are reserved. The remote hyperlinks have been selected for the academic appropriacy of their contents; they were free of offensive and litigious content when selected, and will be periodically checked to have remained so. Copyright © 2003-2018, Derek J. Smith.


First published online 10:35 GMT 11th December 2003, Copyright Derek J. Smith (Chartered Engineer). This version [2.0 - copyright] 09:00 BST 8th July 2018.

Although this paper is reasonably self-contained, it is primarily designed to be read as a subordinate file to our e-paper on "Speech Errors, Speech Production Models, and Speech Pathology".


McCarthy and Warrington's (1984) "A Two-Route Model of Speech Production"

The authors begin with a description of the speech production deficits typically seen in aphasic patients, and then home in on the "two route" model of speech production first clearly expounded by Lichtheim (1885). With reference to the layout of Lichtheim's famous "house" diagram, and, using Lichtheim's own shorthand, the two routes in question are the B-M route, which conveys the main flow of willed and meaningful language production (failure of which results in what is known as a "transcortical motor aphasia"), and the A-M route, which generally bypasses the higher conceptual system, and is used in simple repetition tasks (failure of which results in what is known as a "conduction aphasia"). The authors then report two cases "in whom there was a very marked deficit in speech production in repetition tasks but whose spontaneous speech was relatively well preserved" (p464), and another "in whom the opposite pattern of deficits was observed". Here are the relevant case details .....

Case 1 - ORF, Conduction Aphasic: ORF was a 59-year-old left-handed farmer, who had suffered a right parieto-temporal haemorrhagic stroke in November 1981. After three weeks his verbal IQ was recorded at 99, and his vocabulary and comprehension were generally good, as was his spontaneous speech. However, he performed extremely poorly on the digit span test, and "was quite unable to repeat three digits forwards" (p465). He also performed weakly on naming-from-description tasks, and had a marked deficiency" in repeating words. The authors summarise his deficits as "typically associated with dysfunction of the posterior areas of the dominant hemisphere".

Case 2 - RAN, Conduction Aphasic: RAN was a 51-year-old right-handed plumber, who had suffered a left parietal thrombotic stroke in January 1982. When tested after five weeks, he displayed "clear word-finding difficulties" but made negligible phonemic errors. He also suffered "a severely impaired short-term memory and a speech production deficit in repeating single polysyllabic words in the context of relatively preserved spontaneous speech" (p467). 

Case 3 - ART, Transcortical Motor Aphasic: ART was a 58-year-old right-handed school lecturer, who had suffered a left parietal infarct in August 1982. When tested over the next three months, he displayed severe word finding difficulties, and had a "marked speech production deficit". "His comprehension of written and spoken single words was excellent and his ability to repeat was significantly better than his ability to produce words in reading, naming, or spontaneous speech" (p469).

Guided by these initial profiles, a number of more exploratory tests were then carried out on the three cases (although not all on all), as now summarised .....

Experiments 1 and 6 - Single Word Repetition and Reading: In Experiment 1, ORF and RAN were tested with a carefully selected set of single word stimuli presented (a) for repetition when spoken by the researcher, or (b) for reading out loud from a row-and-column matrix. In Experiment 6, the same tests were carried out on ART. The stimuli varied in frequency of use and syllable length. Here are the results for the repetition test [HF = High-Frequency Words; LF = Low-Frequency Words]. Note the effect of word frequency upon accuracy.


































Experiment 2 - Repeating Nonsense Syllables: When required to repeat nonsense syllables ORF had severe difficulty, but could nonetheless make same/different phonemic judgements from one stimulus to the next. This pattern of impairment suggests that the low frequency problems seen in Experiment 1 above cannot simply be attributed to defective short-term memory.

Experiments 3 and 7 - Repeating Cliché Sentences: There have been many reports of automatic speech being preserved in patients suffering with Broca's type aphasia, back to, and including, Broca's (1861) own patient, Leborgne. To help explain this phenomenon, it is often suggested that clichés - multiple word phrases such as "he put the cart before the horse" - have a single lexical entry despite consisting of a number of separate words. Both ORF and RAN [Experiment 3] were therefore asked to repeat a carefully selected set of 40 clichés, and both scored worse on that sort of material than with normal sentences matched for length and complexity. ORF scored 85% on sentences, against 50% on clichés, and RAN 45% against 18%. When the test was repeated on ART [Experiment 7], he scored 53% on sentences and 73% on clichés, that is to say, the opposite pattern. 

Experiment 4 - Sentence Context: ORF's repetition ability was tested with probe words either (a) as the last word in a natural sentence, or (b) in isolation. Spoken and written versions of the test were administered. Performance on the single word auditory condition was worst (at 40% accuracy), followed by the sentence auditory condition (at 50% accuracy), followed by the two reading conditions (at 65% and 63% respectively).

Experiments 5 and 8 - Enforced Semantic Processing: ORF and RAN were tested with sentence stimuli, each one ending with a low frequency three-syllable test word, and either containing or not containing a "nonsensible" word. The subjects had to carry out two tasks with each sentence, namely (1) to repeat it, and (2) to indicate by nodding or shaking their head whether they thought it made sense or not. The second task thus forced the patients to process the test material at semantic level. Both patients repeated the test word better in the enforced semantic processing condition, suggesting that there are TWO phonological encoding processes, one working on auditory inputs and one working on semantic inputs [as shown in boxes D and C respectively in the figure below].

McCarthy and Warrington conclude as follows .....

"We have thus shown that the locus of the speech production deficits observed in these three cases does not arise in speech perception, comprehension, or short-term storage. We would argue that the deficits lie in the transcoding of information between input and output systems." (p480; italics original)

They then put forward the following diagram as a "description of the double dissociations which we have obtained" (p481) .....

McCarthy and Warrington's (1984) "Two-Route" Speech Production Model: Diagram (a) shows McCarthy and Warrington's "functional model" of the speech production process. This shows two optional routes between an input process A [top, yellow] and an output process E [bottom, mauve]. The left hand route down the diagram transcodes without going through the comprehension process B [top right, green], and damage here results in conduction aphasia (typified by repetition defects). This is therefore the pathway believed to be dysfunctional in ORF and RAN. The right hand route down the diagram understands at B, before transcoding at C, and damage here results in a transcortical motor aphasia (typified by spontaneous speech defects). Technically speaking, Diagram (a) is a five-box box-and-arrow dataflow diagram at about the second level of functional decomposition. The master (or "level zero") diagram would have been cognition as an unanalysed whole, fed by inputs and producing outputs, and the first level of decomposition would have introduced three fundamental sub-processes, namely perceptual processing, higher processing, and motor processing. We may therefore cross-map the five processes shown in Diagram (a) onto the three processes shown in the Lichtheim diagram, in order to show their historical provenance, and we do this in Diagram (b) [Lichtheim's House is the half transparent inner triangle], and what we might therefore (by extension) term "McCarthy and Warrington's House" is the image superimposed thereon].

If this diagram fails to load automatically, it may be accessed separately at


[McCarthy and Warrington's (1984) 'Two Route' Speech Production Model

Diagram (a) redrawn from a black-and-white original in McCarthy and Warrington (1984, p481; Figure 4). Diagram (b) ours. This combination graphic Copyright © 2003, Derek J. Smith. For instructions on how to build technically elegant DFDs, see our e-tutorial on "How to Draw Cognitive Diagrams".