Figure 135

Percentage of writing time devoted to planning (P), sentence generation (SG), and revision (R) as a function of whether an outline had been produced. Based on data in Kellogg (1988).

themes. Children and adults wrote an essay, and thought aloud while planning what to write. Those participants who produced a high-level main point used on average 4.75 different knowledge-transforming processes during planning, whereas those who produced a low-level main point used only 0.23 knowledge-transforming processes.

Are expert writers helped by their use of either rough drafts (an initial version of the text) or outlines (focus on main themes)? Kellogg (1988) asked his participants to learn several facts about a controversy over bussing systems for the handicapped, and then to write a business letter arguing for a particular system. Producing an outline increased the quality of the letter, whereas producing a rough draft did not. The proportion of the time during writing devoted to different processes is shown in Figure 13.5. Outline producers spent more time in sentence generation than did the no-outline participants, but less in planning and reviewing or revising.

What do these results mean? Participants who produced outlines had reduced processing load during the writing of the final text. Planning is the most important and hardest process in writing, and producing an outline has the great advantage that the planning process can be almost complete before starting to write the polished draft.

Word processing

There has been a large increase in the use of word processors in recent years. Is this a good thing? Kellogg and Mueller (1993) compared text produced by word processor and by writing in longhand. For highly experienced users of word processors, the rate at which text was produced and the quality of what was written did not differ between the two conditions. However, word processing impaired production rate and quality of text for less experienced users of word processors.

Kellogg and Mueller (1993) assessed effortfulness of writing by measuring reaction time to an occasional stimulus. Word processing produced more effortful planning and revision than writing in longhand, but there was no difference for sentence generation. Those using word processors were much less likely than those writing in long-hand to make notes (12% vs. 69%, respectively), and this may help to explain the findings.

Kellogg (1994, p. 160) came to the following conclusion: "Studies on writing fluency [speed] and quality generally show no difference between composing on a computer or with pen and paper ...a writer should select whatever tool he or she finds comfortable and useful." What is important is that the writer has access to all the relevant knowledge needed for the particular writing task. The method of writing rarely influences knowledge accessibility, and so should have only minor effects on writing performance.


Many brain-damaged patients who have particular problems with spelling and with writing have been studied. The focus here is on two major issues. First, there has been controversy about the role played by inner speech in the writing process. Second, there have been suggestions that the processes involved in spelling can vary. The notion of different "routes" to the writing of words is considered in the light of the relevant cognitive neuropsychological evidence. Cognitive neuropsychologists have focused only on certain limited aspects of writing, especially those concerned with the spelling of individual words (for reasons, see Chapter 19).

Inner speech and writing

It has often been assumed (e.g., Luria, 1970) that writing depends heavily on inner speech, and that we say words to ourselves before writing them down. This makes intuitive sense, and some spelling mistakes (e.g., "akshun" instead of "action") seem to be based almost entirely on knowledge of word sounds. The key theoretical issue is whether inner speech is essential for writing. One approach is to consider brain-damaged patients having little or no inner speech to see whether they can write effectively, and some relevant studies are discussed here.

Levine, Calvanio, and Popovics (1982) reported on the case of EB. He had suffered a stroke, and his overt and inner speech were practically non-existent. In one task, he was given a target picture and four further pictures. His task was to decide which of the four pictures had a name that rhymed with the name of the target picture. He worked out the names and spellings of most of the pictures, but could not use the sounds of the picture names to perform the task accurately. However, EB's writing skills were largely intact. The quality of his written language can be seen in this account (Levine et al., 1982) of his first memories after his stroke:

Gradually after what seemed days and days, got back enough strength to pull myself up and sit if I held on. I tilted off to the right and had a hard time maintaining my balance.The nurse and doctor and an orderly helped me up then.

Most aphasic patients are severely impaired in both writing and speaking, and so lack the marked discrepancy between writing and speaking skills shown by EB. However, there is a potential difficulty in interpreting the findings from EB. He had extensive experience of speed reading, and claimed to be able to read without processing the sounds of words. Thus, what is true of EB may not be true of other patients or normal individuals.

Further evidence that some brain-damaged patients do not rely on inner speech when engaged in writing comes from the study of patients with jargon aphasia (discussed earlier). Their speech is full of non-words which sound similar to the intended words (e.g., "skut" instead of "scout"; "orstrum" instead of "saucepan"). These errors do not seem to be due to problems of articulation, because patients with jargon aphasia can often pronounce common multi-syllabled words accurately.

If jargon aphasics rely on inner speech while writing, then the kinds of errors present in their pronunciation of words should affect their spelling of the same words. Patient RD (Ellis, Miller, & Sin, 1983) was shown various pictures, and called an elephant an "enelust", a screwdriver a "kistro", and a penguin a "senstenz". His spellings of all these words were accurate, but the superiority of writing over speech that RD displayed may not be generally true of jargon aphasics.


In this section, we will consider the various ways in which we spell words that we hear or see. A sketchmap of what may be involved is shown in Figure 13.6. The main points to be noted are as follows:

There are several routes between hearing a word and spelling it.

• The spelling of known or familiar words involves use of the graphemic output lexicon (a store containing information about the written forms of familar words). Heard words can gain access to the graphemic output lexicon either through the semantic system (which stores word meanings), or through the phonological output lexicon (which provides information about the spoken forms of words). Both routes can be used by normal individuals.

• The spelling of unknown words or non-words cannot involve use of the grapheme output lexicon, because no information about them is available in the lexicon. Spellings are constructed from the spoken or phonemic forms of words by means of phoneme-grapheme conversion, which capitalises on the regularities in the language. This route will produce plausible but incorrect spellings of irregular words (e.g., "yot" for "yacht").

• It is assumed that there are four lexicons involved in language processing (visual input lexicon; auditory input lexicon; phonological output lexicon; and graphemic output lexicon).

Next we consider some of the evidence relating to the major assumptions incorporated in the model.

Phonological dysgraphia

What would we expect to find if a patient could make practically no use of phoneme-grapheme correspondence, but the other spelling components or modules were intact? He or she would be able to spell known words accurately, because their spellings would be accessible in the graphemic output lexicon. However, there would be great problems with unfamiliar words and non-words. The term phonological dysgraphia is applied to patients with these symptoms.

A patient fitting this description is PR, who was studied by Shallice (1981). He (PR rather than Shallice!) had made a good recovery from a stroke. He wrote over 90% of common words correctly to dictation, and was only slightly worse in his spelling of relatively uncommon words. However, he was extremely poor at

Stop Anxiety Attacks

Stop Anxiety Attacks

Here's How You Could End Anxiety and Panic Attacks For Good Prevent Anxiety in Your Golden Years Without Harmful Prescription Drugs. If You Give Me 15 minutes, I Will Show You a Breakthrough That Will Change The Way You Think About Anxiety and Panic Attacks Forever! If you are still suffering because your doctor can't help you, here's some great news...!

Get My Free Ebook

Post a comment