Reading and Writing Systems: Conveying and Archiving Language

I had a really great idea just before falling asleep the other night, but I was too tired to get up and write it down. At least I'm pretty sure it was a great idea, although I just can't remember enough of the details to confirm that it made any sense at all. In fact, I have no idea what the idea was about, and try as I may I can't dredge up even the gist to see if I can reconstruct it for further evaluation.

The solution to the foregoing dilemma, or lost opportunity, is to use a system that was invented more than 6,000 years ago: a written record of picts, graphs, or symbols. The alternative prior to such writing systems was complete reliance on one's fallible memory. Elaborate oral traditions involving extensive practice mitigated these memory failures but were easily overwhelmed if the message to be conveyed was constantly changing or the stakes were high, such as a legal contract with monetary consequences.

Because spoken (and signed) languages are brilliant systems for conveying meaning in all its subtle nuance and complexity, any writing system worth its salt must be capable of archiving the intended message as accurately as possible. Given that most language users have an intact visual system, the most natural form of writing consisted of markings on a surface (rather than tactual stimuli as in Braille). These markings had to be sufficiently generative to capture the complexity of spoken language yet easily encoded by a reader to be efficiently processed without recourse to lengthy scrutiny. Beginning as scratches on parchment, reading is now pixels on a screen and writing is taps on a smartphone.

Given enough practice, the average 10-year-old has done a remarkably good job of learning to read and write. As a result of the near universality of literacy, almost everyone takes for granted that reading is “natural” and “effortless.” Yet when a culture fails to support systems for instructing their children to read, when children are deprived of access to the written word, or they have a neural deficiency in how words are encoded and translated into sounds, we become painfully aware of the fragility and complexity of our ability to read.

The three articles that comprise this special issue on reading, writing, and spelling provide an intriguing overview of many of the fundamental issues that characterize the systems of writing, the mechanisms by which children learn to read and write, and the brain mechanisms that support these systems. Perfetti and Harris (this issue) provide a nice overview of the different types of writing systems (i.e., logographic, syllabic, alphabetic) and make a key point that is reiterated in the other two articles: every writing system evolved from and is intrinsically linked to a spoken language. Thus, although in principle one could design a writing system that conveys meanings “directly” (i.e., without recourse to spoken language), that is not how any writing system works in practice—even logographic writing systems such as Chinese, that do not employ written elements to represent syllables or phonemes, nevertheless engage sound-based processes. Perfetti and Harris also provide intriguing evidence from brain imaging studies using fMRI that different writing systems activate both language-specific brain areas and language-general brain areas, suggesting that the type of writing system induces or taps into neural specializations, while at the same time relying on general neural systems that evolved for other reasons.

Treiman and Kessler (this issue) focus on alphabetic writing systems and how young children come to represent the correspondences between sounds and letters. Particularly intriguing is the so-called prephonological stage of writing in which children appear to be all over the map in trying to represent words that they know how to speak in their writing. Treiman and Kessler entertain and reject the hypotheses that children in this stage rely on a syllabic representation or are randomly grabbing the most frequent letters in an attempt to spell known words. Rather, children appear to rely on a form of statistical learning in which they implicitly capture regularities in spelling that are instantiated in words to which they are exposed in text. Interestingly, this process appears to be similar to the statistical learning mechanism that plays itself out in spoken language, where toddlers are sensitive to the probabilistic phonotactics of the phonemes (and diphones) in their ambient language input (Chambers, Onishi, & Fisher, 2003). Trieman and Kessler also point out a variety of understandable confusions facing the young speller, such as the uncertainty about consonant clusters (one symbol or two?) and the fact that the “name” of a letter contains a vowel when spoken out loud (e.g., the “ee” sound in the letter B). Thus, the word car could plausibly be spelled “KR” because the letter R is pronounced with a preceding “ah” vowel.

Seidenberg (this issue) continues the scientific themes of the preceding two articles, most notably the concept of “deep” versus “shallow” orthographies (i.e., the complexity of how letters map onto sounds and vice versa) and the intrinsic reliance of writing systems on the child's pre-existing spoken language knowledge. An ironic feature of this deep/shallow distinction is that shallow orthographies are easy to read “out loud” (i.e., to correctly pronounce novel words), but such facility bears little relation to understanding what is being said. Again, we see that reading as a form of communication must be based on the underlying knowledge of the language that is being used to convey meaning.

Seidenberg (this issue) tackles two further issues of considerable import for education and public policy. The educational issue focuses on how best to raise children's reading proficiency. Seidenberg's main point here is that educators emphasize the motivational aspects of reading: if you find reading boring or effortful you will not read, and failing to read will dampen your reading proficiency. While this is certainly true, reading and writing without feedback is not sufficient to guarantee proficiency. The science of reading emphasizes that language use, in all its varied forms, serves as the scaffold from which reading as a skill develops. Rich exposure to spoken language, therefore, is what can be acquired with little instruction (see Hurtado, Marchman, & Fernald, 2008, for effects of variation in language exposure on lexical processing), whereas skill in reading requires instruction about the systems of (seemingly arbitrary) rules and their exceptions as well as extensive exposure to a diversity of written texts.

Seidenberg's second issue concerns the achievement gap in reading proficiency among African-American (AA) children. His main point here is that the spoken dialect of AA English has no corresponding written form; thus, AA children are confronted with two spoken dialects (i.e., AA English and Standard American English) that map onto a single written form. This places an additional burden on AA children who must “switch” their sound-letter mapping system when reading and writing in one of their two dialects. This dilemma is interestingly different from being a “true” bilingual because having two separate writing systems, each associated with a different spoken language, is easier to deal with than having one writing system shared by two spoken languages.

A final point that each of the three articles makes is the speed with which writing systems change. There are a variety of factors, including cultural norms, ease of spoken dialect shifts versus difficulty of altering written forms given their archival nature, richness of context in spoken language versus subtle/opaque use of context in written language, ease of learning alternative spellings, and/or exceptions to sound-orthography rules, that make any simple characterization of the “optimal” writing system unknowable. What is clear from these three articles is that reading, writing, and spelling are essential skills that require (a) extensive practice, (b) a rich spoken language system, and (c) mastery of a set of mapping rules (with exceptions) that are acquired over a much more protracted period of development than spoken language. Given the complexity of these systems and how they are mediated by different brain mechanisms, reading will remain fertile ground for scientific study as well as a topic of considerable importance for education and public policy for many years.

Acknowledgments

This article is an introduction to a set of three articles with their origins in an invited symposium on reading held at the annual Society for Language Development meeting, November 1, 2012, at Boston University. Preparation was supported, in part, by a research grant from NIH (HD-037082).