Musings about language and literacy and learning

The Relation of Speech to Reading and Writing

OK, we’re here, at our third paper in our series examining the naturalness, or not, of gaining literacy.

  1. Speech is universal. Many languages don’t even have a written form.
  2. Speech has been around far longer than written language.
  3. For each of us individually, speech develops far earlier than reading and writing (if we are fortunate to even develop reading and writing).
  4. Speech does not need to be taught; it is pre-cognitive, like seeing and hearing. Literacy is rather an intellectual achievement.
    • I paused on the first part of this claim. For students with a developmental language disorder, language does need to be taught more intentionally and supported more intensively. And for the type of language that is not just everyday social language—disciplinary, academic written language—such language also needs to be taught explicitly, most especially for multilingual learners, and its acquisition certainly represents an intellectual achievement!
  5. Parts of our brain have evolved to be utilized specifically for language, while reading and writing must both exploit those innate aspects along repurposing other (originally) nonlinguistic parts. This is the “bootstrapping” notion many more current reading researchers speak to based on brain scans (Wolf, Dehaene, etc).
  6. This one is kinda hard to summarize, but it’s basically centered around the idea that writing systems are both constrained by the oral language they are based on, and more variable. Scripts cannot be purely sound based, as speech is — instead, they are “pitched at the more abstract phonological and morphophonological levels” and this greater abstraction requires greater conscious awareness, at least initially, on the part of the learner.
  7. “Speech is the product of biological evolution, while writing systems are artifacts” — “part discovery, part invention.” Here, Liberman echoes an important point also made by Gough and Hillinger:

“The discovery—surely one of the most momentous of all time—was that words do not differ from one another holistically, but rather by the particular arrangement of a small inventory of the meaningless units they comprise. The invention was simply the notion that if each of these units were to be represented by a distinctive optical shape, then everyone could read and write, provided he knew the language and was conscious of the internal phonological structure of its words.” [bold added]

Here’s the similar quote from G&H:

“Whether recognition of individual letters causes difficulty or not, the recognition that each ciphertext word is composed of a sequence of meaningless elements must be hard for the child to achieve. The requirement that he note the same fact about the plaintext, that he recognize that each spoken word is composed of a sequence of meaningless elements, may be even more unnatural.” [bold added]

Gough, P. B., & Hillinger, M. L. (1980). Learning to read: An unnatural act. Bulletin of the Orton Society, 30, 179–196.

This point, made both by G&H and Liberman, is worth pausing on and amplifying in more depth, because it’s not only a key point of departure from the argument of the Goodmans, but furthermore a key point that underlays debates about phonics even today. For the Goodmans, as with many phonics critics since, the point of reading instruction should be that it is facilitated by learning focused on meaning. According to the Goodmans, when a teacher explicitly and sequentially teaches the meaningless, artificial components of phonemes and graphemes, they create a barrier to natural learning:

“With the focus on learning, the teacher must understand and deal with language and language learning. . . . The learners keep their minds on meaning. . . The crucial relationships of language with meaning and with the context that makes language meaningful is also vital. . . .We must focus more and more attention on how written language is used in society because it is through the relevant use of language that children will learn it. They will learn it because it will have meaning and purpose to them.

With the focus on teaching both teachers and learners are dealing with language often in abstract bits and pieces. . . . it’s a serious mistake to create curricula based on artificial skill sequences and hierarchies derived from such studies.

Our research has convinced us that the skills displayed by the proficient reader derive from the meaningful use of written language and that sequential instruction in those skills is as pointless and fruitless as instruction in the skills of a proficient listener would be to teach infants to comprehend speech.”

Goodman, K. S., & Goodman, Y. M. (1976). Learning to Read is Natural.

For the Goodmans and many proponents of balanced literacy today, a focus on meaningless, unnatural components is an impediment to the naturally motivated learning of children. And hey, they aren’t wrong — learning these abstract aspects of oral and written language is a barrier to all too many.

But we should be absolutely clear that Gough, Hillinger, Liberman, and many researchers focused on literacy fully acknowledge that these components are difficult and a tremendous potential barrier to learning—in fact, they fully agree with the Goodmans that learning sublexical units (phonemes and their haphazard letter sequences) is unnatural! The key difference is that they also argue that these artificial units are essential to reading and therefore must be tackled head on and overcome by children in order for reading to truly be successful.

But I’m taking us away from Liberman, and he’s only getting started. He takes some time to outline what he calls “the conventional view of speech.” According to Liberman, this is a view that assumes that speech is governed by general motor and perceptual systems, rather than ones specialized for language. This means that the processing of speech must therefore be cognitive in nature, as it requires translation–similar to learning the written form, it requires attaching a phonetic label to the sounds of what is heard. In this sense, then, learning language can be perceived as biologically secondary.

The reason Liberman takes time laying this out is because if we are to take this view seriously, it means we must see written language as “equally natural” to speech because it is essentially a similar process of coding that requires cognition, with the only difference being one of mode.

This is pretty much exactly what the Goodmans argued: “. . . if written language can perform the functions of language it must be language.”

The other big issue with the “conventional” view, according to Liberman, is that it means “the elements of a writing system can only be defined as optical shapes. . . [and] makes it hard to avoid the assumption that the trouble with the dyslexic must be in the visual system.” This mistaken assumption is indeed a continuing confusion for many about learning to read, as witnessed by some who attempt to teach kids to read by noticing the shapes of words (a quick aside for some nuance: some with dyslexia may have visual-spatial issues, which may become more apparent when learning non-alphabetic written languages, such as Chinese).

word shapes

Here Liberman makes a key distinction: the evolution of oral language is biological, while written language is cultural. I find myself both deeply compelled by this claim, as it is useful, and also a little resistant. I resist because language is also clearly cultural. But I get that the point here is that the mechanism for learning language is baked into our brains, developing rapidly even as we are in the womb, while acquiring literacy is more dependent on cultural transmission and a significant amount of work.

“In the development of writing systems, the answer is simple and beyond dispute: parity was established by agreement. Thus, all who use an alphabet are parties to a compact that prescribes just which optical shapes are to be taken as symbols for which phonological units, the association of the one with the other having been determined *arbitrarily. Indeed, this is what it means to say that writing systems are artifacts, and that the child’s learning the linguistic significance of the characters of the script is a cognitive activity.” [bold added]

This leads Liberman to propose what he calls the “unconventional view of speech.” I’m going to do some heavy paraphrasing here, but if you’re into speech pathology or like to geek out about the articulatory dimensions of speech, you may find this section of the paper interesting, as he lays out why “co-articulation” is a fundamental aspect of speech. Essentially, he lays out some principles that allows for the claim that “There is no need . . . for a cognitive translation from an initial auditory representation, simply because there is no initial auditory representation,” meaning that speech is processed rapidly and naturally.

And now Liberman turns to the Goodmans directly to take their full argument head on, so it’s worth reproducing this section in full:



“The conventional view of speech provides no basis for asking this question, since there exists, on this view, no difference in naturalness. It is perhaps for this reason that the (probably) most widely held theory of reading in the United States explicitly takes as its premise that reading and writing are, or at least can be, as natural and easy as speech (Goodman & Goodman, 1979). According to this theory, called ‘whole language,’ reading and writing prove to be difficult only because teachers burden children with what the theorists call bite-size abstract chunks of language such as words, syllables, and phonemes’ (Goodman, 1986). If teachers were to teach children to read and write the way they were (presumably) taught to speak, then there would be no problem.

But if we adopt the “unconventional view ” of speech, then we don’t view spoken and written language, one auditory and the other visual, as equivalents. Instead, this view allows us to see that speech is processed completely differently, and much more swiftly, and we don’t need to become aware of nor think of the sub units of sounds within a word: “there is nothing in the ordinary use of language that requires the speaker/listener to put his attention on them.”

“The consequence is that experience with speech is normally not sufficient to make one consciously aware of the phonological structure of its words, yet it is exactly this awareness that is required of all who would enjoy the advantages of an alphabetic scheme for reading and writing.”

And the specialized properties of speech, such as co-articulation, which allow us to wield and process them so efficiently, actually present us with a greater barrier in conversion to written language. Co-articulation, which is when we merge sounds together in the speech stream, “has the disadvantage from the would-be reader/writer’s point of view that it destroys any simple correspondence between the acoustic segments and phonological segments they convey.”

Thus and therefore, learning to read and write requires cognitive work, at least initially, that is not required for spoken language (Note that though I’ve taken this paper at face value with the word speech and we’re focused on those aspects specific to spoken language, many of these characteristics can apply just as readily to sign language).

Whew! This paper was a bit harder to unpack than the others, but I think it’s a very good capstone to our investigation in the series. So are we convinced that learning to read, at least initially, is unnatural?

I’ll pursue some final thoughts to wrap up some loose ends in the next post.

#natural #unnatural #reading #spokenlanguage #writtenlanguage #language #literacy #speech #meaning #Liberman