Most adults read about 200-250 words per minute (wpm), but Spritz, a reading application that is attracting considerable social media attention, claims that most people can easily double or triple this speed without any special training.
Normally when we read, our eyes move along the lines of a text, landing (fixating) on words for a tenth to a quarter of a second, then making short jumps (saccades) to the next word.
The developers of Spritz claim that, in the traditional method of reading, only 20 per cent of reading time is spent processing the content of a text and 80 per cent is devoted to moving the eyes between words.
Their solution is to eliminate the need to make eye movements. Words are presented one at a time, beginning at the typical reading rate of 200wpm, and the reader is encouraged to gradually increase it to rates of up to 1000wpm.
Sounds good, doesn't it? At that rate, you could read a novel in 90 minutes. But there are some aspects of reading that apps such as Spritz don't quite nail.
The science underlying the Spritz technique relies on two well-established characteristics of eye movements during reading:
• Skilled readers' perceptual span - the window of text we use during reading - is about 13 characters. This is the maximum length of word exposed in the Spritz "redicle".
• We characteristically land our eyes at a predictable position in the word - between the beginning and middle of the word - that Spritz refers to as the optimal recognition point (ORP).
Spritz' main innovation is to centre the word in the redicle on the ORP and highlight it in red. This is claimed to speed up reading by ensuring that the reader fixates at the optimal location to identify the word, while eliminating the time required for the reader to compute this location and move the eyes to it.
Spritz takes almost the opposite approach to increasing reading speed as the "standard" approaches to speed reading spruiked in hundreds of YouTube clips.
These methods assume that sequential word-by-word reading is the main barrier to rapid reading and advocate a variety of methods designed to break this habit and adopt non-sequential scanning strategies, such as moving the eyes down the centre of the page, that are claimed to facilitate unconscious processing of relevant information in the text.
Despite the very different ways in which they aim to achieve it, the methods do, though, have a common goal of reducing subvocalisation - saying the words in your head - during reading. In standard methods, eliminating subvocalisation is focused on in training.
In Spritz, it is an automatic outcome of "spritzing" because the average rate of speech is less than 200wpm. Subvocalisation cannot be maintained at rates higher than that.
On the surface, Spritz is better aligned with scientific evidence about the skilled reading process than standard speed reading methods. Even skilled readers fixate on most of the content words (nouns, verbs, adjectives) in a text, although they often skip over short function words (such as "to", "in", "on", "the") and highly predictable words.
Skilled readers' general strategy is, therefore, more similar to the sequential strategy forced by Spritz than the non-sequential scanning strategies advocated by many standard approaches to increasing reading speed. A sequential reading strategy is also important for comprehension, particularly in English where the order of words is important for meaning.
But at a deeper level, Spritz ignores critical aspects of the scientific evidence about eye movements in reading. Most importantly, it ignores the time and cognitive effort required to integrate the words in a text for comprehension.
Although there is some truth to the claim that the relatively slow pace of eye movements reflects physical constraints on eye movements, it is primarily because of the cognitive demands of word identification and comprehension.
The time we fixate on words depends on their familiarity, predictability and length - the factors that determine the time required to identify and integrate their meanings.
We also pause at clause and sentence boundaries to conduct "wrap up" processes that are important for effective comprehension. Removing readers' control over which words they fixate on and how long they look at them reduces comprehension.
Systematic research conducted in the 1970s investigating "rapid serial visual presentation" (RSVP) methods that present text one word at a time found that comprehension fell rapidly beyond rates of about 500wpm, particularly for texts longer than single sentences.
The Spritz developers' assertion that retention levels are at least as high as for traditional reading requires more detail to convincingly demonstrate that using the ORP overcomes these limits on comprehension.
Essentially, Spritz forces people to process written language like speech - one word at a time with no opportunity to go back to check any errors in word identification or interpretation, as we do quite frequently during normal reading.
Obviously, we are very effective at understanding speech, and can apply those same skills to spritzing.
But speech contains a range of additional cues, such as intonation, pauses and gestures, which all contribute to comprehension.
Speech is also usually simpler than written language and focused on well-defined topics, reducing the demands on working memory associated with its sequential presentation.
Most critically, the typical rate of speech is around 200wpm. The convergence with the typical rate of reading may be accidental, but most cognitive scientists would attribute the similarity to the bottleneck caused by the attention and memory processes required for comprehension in both modalities.
These concerns about comprehension may be of little relevance for the social media applications that Spritz is designed for. Such content may be closer to spoken than written language in its complexity.
Spritzing may be an effective delivery mode for tweets of less than 140 characters and for small-screen devices where there is little opportunity for readers to scan text. However, the need for users to stare even more fixedly at the middle of a screen may exacerbate the antisocial impact of such devices.
Will spritzing yield transferable skills that benefit reading of standard text?
The claims about extraordinary increases in reading speed with training in standard approaches to speed reading have not survived scientific scrutiny, but the skimming strategies they teach are useful in many reading contexts.
Perhaps similar benefits will follow from Spritz users discovering that they can understand text without "saying the words in their head". This may encourage the use of more flexible strategies in "normal" reading contexts - but Spritz reinforces a sequential approach to reading that is incompatible with the flexible, meaning-guided scanning strategy needed for effective skimming.
Perhaps most frighteningly for a reading researcher - and reader - like me, the speech-like processing encouraged by Spritz might contribute to our evolution towards the world envisaged in Spike Jonze's recent film Her, in which written text has become an anachronism.
Deprived of exposure to text, readers may gradually lose the sensitivity to the structure of written language that underlies our capacity to locate the ORP for words and capitalise on the multiple cues in written text that contribute to effective comprehension. But maybe I am just revealing my age - or smartphone envy.
Sally Andrews is a professor of cognitive psychology at the University of Sydney.