Horror fans not familiar with his work should check out the movies of director David Cronenberg, including the 1981 creep-out Scanners.
It is commonly said (by stutterers and others) that while speaking, a stutterer will scan ahead to look out for 'difficult' words. An Internet search for [stutter + "scan ahead"] will verify the popularity of the idea. The word scan, taken literally, implies looking. We scan text on a page. We are able to do so because the text is, in fact, already on the page and available to be scanned. While we can take the word scan figuratively, as might be argued by some, the fact is that the word scan implies sight, and sight implies written text. This sort of confusion between spoken and written language shows up repeatedly in the stuttering literature. I'd like to question the degree to which such scanning actually occurs, and to what degree the analogy to a visual process (scanning) obscures rather then elucidates whatever it is that is actually happening.
First, let's look at the claim. Upon recognizing a difficult word coming up, the stutterer may insert a synonym, rephrase to avoid the word, or start an entirely new sentence in mid-speech. We need not rely on the word of stutterers for such mid-sentence changes. A recording of a stutterer talking studied closely will often reveal a pause, mid-sentence, followed by a grammatical mash-up that produces a new finish to the sentence that is divorced from what came before. A very similar kind of sentence breakdown is actually common in the casual speech of non-stutterers, and is often unnoticed by the listener. Transcripts of the extemporaneous speech of politicians that include such verbal inconsistencies have, however, been used (unfairly) to embarrass them by their opponents in the media. What is ignored (and acceptable) during speech becomes obvious on the page.
The fact this sort of word avoidance and sentence restructuring by stutterers occurs does not, however, necessarily support the idea of scanning. That is, if the decision to avoid a 'difficult' word doesn't come until the word is arrived at, then there was no scanning. Scanning implies looking ahead multiple words in a serial sense, and the behaviors of word replacement or sentence restructuring does not prove that scanning has occurred.
So how far 'down the page' (figuratively) can a speaker scan when he or she doesn't normally know exactly what they will say until they say it? Speech is a subtle and complex process by which vague, unformed ideas become words in grammatically ordered sentences at a rapid rate without conscious effort. So how much scanning of upcoming speech can a stutterer actually do?
Surely we cannot imagine that anyone can know in any but the rarest cases what is coming two sentences ahead during spontaneous conversation. The best one can say for words outside the current sentence is that if one is discussing Abraham Lincoln, there will be a time when one is going to have to say the name Abraham Lincoln, the word President, and other such terms necessary to the topic. We may perceive that while we speak on sentence, the next sentence will logically have to include Lincoln's name, or 'the President,' but that is not 'scanning,' in any non-stuttering sense of the word.
Putting aside such cases in which it becomes obvious that a word will have to be spoken some time soon, what can we say about the realistic possibility of 'scanning' ahead in speech? When we begin a spontaneous sentence, we do not have in our conscious mind how it will end. Sentence construction is done on a subconscious level, while we observe our listener, wait for the coffee machine to finish cycling, and wonder what that unpleasant smell is in the office lunch room.
The actual work of scanning, so-called, can only occur when a decision has been made to use that word in a particular sentence, in a particular grammatical construct. And given that in spontaneous speech one sentence cannot be logically constructed until the previous sentence is already known, we have to assume that the horizon for knowing upcoming words is very short.
What it comes down to is that it is difficult to imagine how one could scan for a word in the second half of a sentence until already engaged in speaking the first half. One could say that we need to know where our subconscious speech production center is taking us before we can know what our next steps will be.
Let's look at some examples. In a simple sentence of a single clause, the final words of the sentence may come to our conscious mind after we have said the first few words. This is not a matter of scanning ahead; we simple know where the logic of the sentence is going before we finish speaking it.
In a longer, more complex sentence, we may perceive the words that will begin a second clause as we approach speaking the end of a first clause. This may also be true of sentences. We may know as we finish one sentence how the next sentence will begin. This is less a matter of scanning that of gaining consciousness of the next few words in a package immediately before we speak them. We could make an analogy to walking in the woods. We do not scan up the path to prepare for rocks and roots on the ground. Rather, we keep an eye on the path immediately in front of us, and adjust our footfalls only as we reach the obstacle.
We can now ask whether this process, whatever we call it, is unique in any way to stutterers? Does it really require a life-long experience with a speech pathology to be occasionally conscious of the next few words to come as we speak? Being a stutterer, I can't speak for non-stutterers, but I suspect the answer is no. A reasonable assumption is that everyone becomes aware of upcoming words as they become 'next in line' for speech. For the normal speaker, this awareness is ephemeral and trivial. They might use it to edit themselves, but the operation would be a barely conscious one.
To sum up: I believe that what is being described by the term 'scanning ahead' is not scanning at all. The fear of saying certain words - and of stuttering while doing so - is not scanning. The consciousness of immediately upcoming words during speech could only be called scanning by stretching the word to its breaking point. A page of text can be scanned down a few (several?) lines while reading. Paragraphs of speech yet to be formed in the mind cannot be scanned under any effort of the imagination. The knowledge that certain words are associated with particular subject is not scanning - it is a static state rather than an active process. So while the term 'scanning ahead' points to a very real phenomenon - an effort of circumlocution to avoid a stutter block - the word can only suggest to non-stutterers something far beyond what actually occurs.
And why does it matter? It matters because this field is so full of obscurantist language that it is difficult to discuss the pathology using the standard jargon without necessarily being led astray by the very terms themselves. Word fear is very real, and can interfere with communication when it results in meaning-destructive circumlocutions. And that process of generating any such circumlocutions can, in itself, reinforce the fear that began the cycle.
I would suggest that rather than scanning ahead for difficult words, stutterers are simply alert to the occurrence of such words in their speech as they bubble up in consciousness. Anxious alertness would be a better descriptor than the 'scanning ahead' cliche found so often in the literature of stutter.