Tuesday 30 October 2018

How fast could CS Lewis read? A guest post from Kevin McCall

Note: This was originally published at Superversive Inklings

It is well known that C.S. Lewis was an extremely fast reader. Richard Ladborough, in his essay “In Cambridge” in the book C.S. Lewis at the Breakfast Table writes: “It is now common knowledge that his [Lewis’s] memory was prodigious and that he seemed to have read everything.”

 In his essay “Jack on Holiday” in the same book, George Sayer says, “But when the coffee or tea had been cleared away (I think he preferred tea), he liked to settle down to an hour or two of silent reading. He would choose a book from my shelves, usually a novel, and often one that he had read before, for he held the view that the qualities of a good book could not be appreciated at the first reading. … He read very fast and if the book were a humorous one (he pronounced that word with an h) often chucked or laughed aloud.”

In his essay “C.S. Lewis: Supervisor” collected in C.S. Lewis Remembered, Alastair Fowler compares his own reading to Lewis’s: “Reading habits, of course, were different in the fifties; I used then to read ten hours a day. Lewis, who read far faster, read with surer grasp, and read whatever commitments allowed – read even at mealtimes – read prodigiously.” Lewis also completed an English degree at Oxford (with First Class Honours) within one year rather than the typical three years.

Not only was Lewis a fast reader, he also had an extraordinary memory for the details of what he had read. Douglas Gresham writes in his foreward to A Grief Observed “Helen Joy Gresham (née Davidman), the ‘H.’ referred to in this book, was perhaps the only woman whom Jack ever met who was his intellectual equal and also as well-read and widely educated as he was himself. They shared another common factor: they were both possessed of total recall. Jack never forgot anything he had read, and neither did she.”

Alastair Fowler also recalls Lewis’s powerful memory: “The flow of Lewis’s writing and speaking had much to do with this remarkable memory. Memory feats were common enough in Oxford then, especially among classicists. Edgar Lobel the papyrologist and fungiphage, to mention one, modestly denied having Homer by heart – but added, ‘Mind you, if you said a verse I dare say I could give you the next one.’ Lewis could have claimed much the same of Paradise Lost.”

It is natural to ask, how fast could Lewis read? Also, how did Lewis read? Was he an efficient skimmer who knew which words to skip or did he have some nearly superhuman ability to take in huge volumes of text? A letter written to C.S. Lewis’s brother Warren Lewis on May 10, 1921 provides a way to calculate Lewis’s reading speed. Here is the relevant passage: “Which reminds me, did you ever read Daudet’s ‘L’Immortel’? It is a novel about the Academie Francaise: if you like sheer cool premeditated insolence you should order this by the next mail – tho’ perhaps I should warn you that it is only a couple of hours reading, and you may like books that last, on the world’s end.”

Copying the Project Gutenberg English translation of L’Immortel into a word document shows that the book contains 69,265 words. Rounding to 70,000 and dividing by 120 minutes indicates that Lewis read 583 1/3 words per minute (Wpm). If we assume a single-spaced page is approximately 500 words, Lewis could read 70 pages per hour. (Though if Lewis was reading the book in French, his English reading speed is likely faster).

Ronald Carver’s 1985 paper in Reading Research Quarterly, “How good are some of the world’s best readers?” provides some clues to answer the second question. Carver selected 16 readers on the basis of their excellent reading ability: four college students who made high scores on a test of vocabulary and reading comprehension, four accomplished speed readers, four professionals whose jobs required large amounts of reading (“a writer for the New Yorker magazine,” “a copy editor for a major metropolitan newspaper who had been recommended by managing editor as one of the best of the 12 copy editors they employed,” “the former head of a major medical school who had served as editor of a nationally known medical journal,” “a history professor at a major university who also wrote book reviews for newspapers”), and four people who had achieved very high scores on various tests (a member of Mensa who had been successful on the same tests given to the college students, a perfect scorer on the SAT, a perfect scorer on the GRE, and a test taker who had scored above 700 on the GMAT). Carver gave tests of reading comprehension, vocabulary, and intelligence (Raven Advanced Progressive Matrices). Most importantly, Carver tested subjects’ ability to write summaries of various 6,000 word passages and to recall important details from these same passages when given time to read the passage corresponding to reading speeds of 24,000 Wpm, 6,000 Wpm, 1,500 Wpm, and 375 Wpm.

Carver found that while many of the research participants could write good summaries while only spending a short amount of time with the passages, their ability to recall details fell precipitously as time given to read passages decreased. He concluded that none of the readers showed the ability to recall and comprehend details taken from the entire passage above while reading at speeds above 300-600 Wpm. Beyond 600 Wpm, the speed readers were really skimming, not reading.

Carver describes reading and skimming as follows: “What is ordinarily called reading involves an attempt to comprehend the thoughts the author intended to communicate on a sentence-by-sentence basis … When skimming, the individual does not attempt to comprehend the complete thought expressed in each sentence. Instead, the individual is simply trying to extract as much general information as possible about the passage by sampling only isolated words and phrases.”

C.S. Lewis’s calculated reading speed of 583 1/3 Wpm is at the very top of the range described by Carver, suggesting that Lewis really was reading rather than skimming. Lewis’s ability to recall details with great accuracy provides further support for this view. In addition to Lewis’s rapid reading and excellent memory, he also had profound insight into what he had read and skill in describing it. Alastair Fowler wrote: “For he talked like an angel. My idea of how angels might talk derives from Lewis. His prose is brilliant, amusing, intimate, cogent; but his talk was of a superior order. It combined fluent, informal progression with the most articulate syntax, as if, somehow, it was a text remembered – and remembered perfectly”

Since Carver only tested 16 people, it is certainly possible that an extraordinary reader slipped through his selection process. Among individual readers, the best overall performance was achieved perfect scorer on the GRE, called TEST-GRE by Carver. Carver writes, “Of the 16 superior readers tested, TEST-GRE seems easily to qualify as the best reader in terms of being able to comprehend the most while reading the fastest. There is evidence that she or he could read eighth-grade material at around 500 Wpm.” But, TEST-GRE was randomly selected from among over 200,000 people who had made a perfect score within recent years, so this person could just as easily not have been selected. Similarly, there could have been an even better reader with a perfect score on the GRE who was not selected.

Carver also makes an interesting comment on the fourth best overall scorer, a speed reader selected on the basis of a reported ability to read 81,000 words per minute, “From the Raven Test, and from the scores on the two book tests at 1,500 Wpm, it appears that SPEED-81,000 is an exceptionally intelligent person who ordinarily skims at very fast rates. However, there were no data which replicated the 81,000 Wpm reported when she or he completed the speed-reading course given by the Reading Foundation of California.”

Can we find an upper bound on claims of extremely fast reading? In fact, the well-known savant Kim Peek provides such an upper bound. Peek was able to read two pages of a book in 8-10 seconds (one with each eye) and recall every word he had read with nearly perfect accuracy. Peek could thus read 12 pages every minute, which works out to 720 pages per hour, and assuming again that a single-spaced page is approximately 500 words implies that Kim Peek could read 60,000 words per minute.

However, Peek was not actually reading. He was not comprehending the thoughts expressed on each page but memorizing the text contained in the book. He would have scored almost perfectly on Carver’s tests of detail recall but would not have been able to write a good summary. Peek’s reading speed provides reason to be skeptical of claims to read above 720 pages per hour.


Wurmbrand said...

It would be interesting to know about the conditions under which Lewis typically read. His prodigious reading ability bespeaks excellent concentration. Was that concentration assisted by usually quiet circumstances -- no radio, loud voices nearby, etc.? I suppose that it was. He had not grown up with the distraction of television and digital media, nor, I suppose, even radio. He seems usually to have lived in quiet neighbourhoods, doesn't he? Perhaps he usually didn't have to hear barking dogs, traffic, &c. The people near at hand would usually have been ones who respected reading -- at Little Lea, at Kirk's, in college rooms -- although in The Kilns, during Mrs. Moore's many years there, Lewis would have been liable to experience interruptions in the form of requests to walk the dog and so on. But I imagine that she didn't simply enter his study all the time and chatter.

I don't mean that favourable circumstances -account- for Lewis's superlative reading ability, only that they (if they existed) must have assisted the abilities that he already possessed.

It would be interesting to know how he learned to read. In the United States, many children have been taught by "sight reading," which some argue contributes to reading difficulties that are less common for children who were taught by "phonics." Certainly Lewis was used to reading works with adult vocabulary early on, and many children can do so if they want to. One of my daughters read War and Peace in the Maude translation (I think it was), The Woman in White, Jane Eyre, Pride and Prejudice, etc. by around her 13th birthday because she wanted to.

Dale Nelson

Margaret Dean said...

Along with his rapid reading speed, it's also interesting to note that Lewis enjoyed being read to -- hence the Inklings' usual practice of reading works-in-progress aloud at meetings.

Kevin McCall said...


All excellent points. I think you are exactly right that the lack of distractions not only helped Lewis focus while reading but allowed him the time to spend hours reading.
In his essay about Lewis, Neville Coghill mentioned that while they were studying literature, he and Lewis read 8 - 10 hours per day (which now would be inconceivable for some people now (though many would not balk at spending 8 - 10 hours repeatedly checking social media!)).

I think Lewis may have used a combination of sight reading and sounding out words. Even auctioneers can only speak at 250 words per minute so above that speed reading purely by silently sounding out words would be too fast to comprehend easily.

Hrothgar said...

I don't find Carver's results particularly impressive or enlightening, to be honest. Perhaps in selecting his test subjects he might have done better if he had considered that the most proficient modern students (from which proficient test-takers are not worth distinguishing in this context) do not necessarily (in fact usually do not) represent the finest intellects nor the greatest raw intelligence within a particular cohort, so much as they represent the most able cohort of those who are most willing to subject themselves to the largely intellect-quashing thought codes of standardized, systemized, bureaucratized modern education, and prosper under them.

Much the same can be said of his other subjects. High achievement within the relatively corrupted systems of modern academia and the media more obviously corresponds with a positive lack of moral qualities than the possession of exceptional intellectual qualities (the degree of the latter's correspondence with raw reading speed being in any case unexamined, and open to question on a number of fundamentals). What his results more obviously said to me is that you are not likely to discover exceptional individuals by selecting for success within a system that does not test for the particular factors being sought, and is in general more oriented to the suppression of exceptionality than otherwise; and that no one should really be surprised by this.

Lest it be thought that I am merely pontificating on something of which I know no better, I will counter his findings for now with my doubtless very unscientific survey of Persons Personally Known To Me (though this actually encompasses a greater number of people than his study) and see whether this sheds any light on the subject:

There are at least three people within my own family, considering nearly 20 relatives removed by no more than two degrees of kinship (I have a large family by modern standards), who, so far as I can judge, read at least as well as his highest-scoring test subject. This is to say that they routinely read for pleasure (thus skimming can reasonably be discounted), at a speed of 500 words per minute or greater over extended periods of time, with at least an 8th grade level of comprehension (in practice normally higher).

One of these individuals is my own good self; another is my mother; the last is my elder sister. A now deceased brother combined near-genius level intelligence with mental fragility that eventually led to a complete breakdown (diagnosed as schizophrenia). He was very possibly a fast reader - but not necessarily, as his abilities were more mathematical than verbal, so I will discount him for now.

My mother and sister are both extremely visually oriented people – my sister perhaps slightly less so – and seem to be correspondingly less oriented to matters auditory. For example: Neither can sing in tune or has evinced any particular musical skill - my sister has tried to learn instruments without much success, while my mother was effectively banned from singing at school due to being so atrocious that she was throwing everyone else off key, even by school choir standards. They tend to get impatient when listening to anyone speaking for too long, sometimes have trouble recognizing people by their voices, and (especially my mother) are easily confused by purely verbal explanations. And so on.

Hrothgar said...


I don’t believe this this visual preference is essential in itself, as I am significantly less visually oriented than either of them, and am also more interested in (and able at) music. I do think, though, that it may relate to the techniques naturally utilized by faster readers, which I have reason to believe are primarily visual, and so, for instance, not likely to be aided by early focus on phonics. This is to say that due to the techniques of faster reading being primarily visual and significantly different from phonics-based techniques, people with a strong visual preference also have a greater disposition to use them (but not a monopoly on doing so).

One thing my sister was very sure of when I asked her how she actually went about the process of reading is that she did not read the page line-by-line, going from the end of one line to the beginning of the next. She thought that she focussed her gaze somewhere in the centre of the page, and then scanned the text from left-to-right to take in the rest of the words while moving down the page. My mother reported similar experiences. I am fairly sure now that this is the main reason why all three of us read faster than most people apparently do.

This is what I have always had the sense of doing too – reading down the page, and apprehending the text by sentence and paragraph, then moving on once I have mentally processed what had just been read - not following the lines, phrase by phrase, which would seem very cramped, slow and awkward by comparison. I would only read text in the latter way (which I assume to be the default for most people, judging by what I hear in general and have just read here) if it was in a language I was having difficulties with, or if I really needed to re-read a particular line closely to fully understand what was being said.

I tend to read verse more slowly than prose, and with frequent pauses to sound it out mentally (or even to speak it aloud if it strikes me as particularly fine), but will probably go back to reading it much as I read prose if I hit an dull or aesthetically unpleasing passage.

I do find it much harder to read in my normal way when dealing with onscreen text – more of a conscious effort, at least somewhat slower, and exhausting after any length of time, so I prefer to read actual books if reading anything lengthy, or for pleasure. I suspect this is because my eyes have to move much further to do the same work due to the larger text resolution and the screen being wider than a page; it takes longer for me to apprehend a section of text because of this; and all the slowness and effort demands more concentration, causes more physical strain, and causes me to tire quickly. Scanning text onscreen is not really any more difficult though, being a fundamentally different type of process.

The speed reported for Lewis sounds to me quite reasonable for reading a relatively light novel in a foreign language (in which, however, one is fully fluent, as I assume he was), and is probably quite close to an upper limit for such reading (except perhaps for a reader who is completely bilingual). I’m fairly sure, however, that I personally would read faster than this while reading something non-complex (or at least, containing fluent or straightforward syntax and easily understood ideas) in my native language, while not subject to any kind of interruptions or lapses in concentration. I will assume this must have held true for Lewis too. Somewhere beteen 700-800 words per minute, maybe slightly more, is probably closer to an upper limit for real reading of material that a reader finds relatively easy (what this is, of course, will significantly vary between individuals).

The extremes of savant text-processing capability are so different to what is normally understood when we talk about reading, and so aberrant in terms of normal human functioning and expectations, that I think we need to be clear it is something else entirely, and should be referred to as such.