Stenography and Orality in Dickens: Rethinking the Phonographic Myth

Drawing on Steven Marcus's claim that by learning and practicing stenography in the law courts Dickens had essentially become a "written recording device for the human voice," Ivan Kreilkamp has argued that Dickens brought the "phonographic innovations in voice writing" to the writing of the novel. The difficulty with this argument is that Dickens learned shorthand from a hybrid system—Thomas Gurney's Brachygraphy—that was radically different from the classic phonography of Isaac Pitman's Stenographic Shorthand. Unlike the Pitman system, which linked shorthand symbols directly to sound, the Gurney system mediated the link through letters—the learner had to memorize symbols which stood for letters rather than for sounds. This essay will argue that Brachygraphy's extra level of alphabetical mediation meant that Gurney shorthand was essentially, and unusually, a creative stenographic system. The nature of the creative language processing implicit in the learning of Gurney shorthand will be described and its implications for Dickens's writing processes will be discussed, drawing on examples which suggest that Gurney stenographic processes were themselves represented in Dickens's fiction and involved in episodes from his life. The overall influence of Gurney shorthand on Dickens's language processing suggests that theories regarding his legacy in relation to "orality," particularly his position and role in "phonographic" interpretations of nineteenth-century culture, may have to be reconsidered. At the same time, we should recognize the importance of the Gurney method in influencing Dickens's creative use of language.

Drawing on Steven Marcus's claim that by learning and practising stenography in the law courts, Dickens had essentially become a "written recording device for the human voice" (192), Ivan Kreilkamp has argued that Dickens brought the "phonographic innovations in voice writing" to the writing of the novel.¹ Other critics and biographers echo Kreilkamp's view that Dickens was putting speech on paper phonographically when they label Dickens's representations of character talk as "phonetic spelling."² One important premise of Kreilkamp's phonographic argument is that Dickens was writing words as they sounded and that this form of writing constituted, in Ackroyd's words, "the graphic embodiment of speech" (133). Another is that Dickens was able to transfer voices and sounds into his fictional speech because he had learned to record them in the courtroom with great precision. The practice of careful listening, stenographic note-taking, and reproduction had made him a faithful recorder of what was said in court and this learned habit helped him to transfer the spontaneity of courtroom speech with "virtually stenographic fidelity" (554) into his fictional writing as an authentic representation of the original.³

Although there is an attractive neatness to the idea that the phonographic medium of Victorian stenography facilitated Dicken's ability to transfer the details of real (courtroom) speech into the lively dialogues of Dickensian characters, it fails to take account of the simple fact that the Gurney system of shorthand that Dickens learned was logographic and alphabetical in its design, not phonetic,⁴ and included learning mechanisms that were totally different from those of the phonographic model of Pitman's Stenographic Sound-Hand. It is true, as Gitelman writes, that "phonetic shorthand emphasised the oral character of language" (24) but it is not clear in what way Gurney shorthand is classifiable as phonetic. The aim of this article is to explore the stenographic origins of Dickens's fictional speech by analyzing the differences between the Gurney and Pitman systems, highlighting the different ways in which they would have affected their users' awareness of script in relation to sound. I argue that the difficulties and internal contradictions of the Gurney system, described so graphically in David Copperfield, combined to produce a unique mechanism for the processing of language and that the repeated use of this mechanism provided Dickens with a unique blend of phonotactic and alphabetical creativity—a creativity that the architecture of the transparently phonographic Pitman system was not designed to produce.

If Steven Marcus is correct in saying that learning shorthand made Dickens aware of a different way of representing speech graphically,⁵ it is also true that there are considerable differences between Gurney and Pitman shorthand as writing systems and that the learning of each system would have provided Dickens with a different kind of awareness of the relationship between script and speech. In order to understand what these "different awarenesses" were we need to understand the architecture of the respective systems: "I bought an approved scheme of the noble art and mystery of stenography (which cost me ten and sixpence); and plunged into a sea of perplexity that brought me, in a few weeks, to the confines of distraction" (Dickens, DC, 503–04, ch. 38). Dickens learned shorthand in or around 1828 using a manual called Brachygraphy. Thomas Gurney first edited and published Brachygraphy in 1772, and his son Joseph subsequently revised it into the "improved" version that Dickens would have used. The 15th edition, published in 1825, contains a preface by William Gurney, the founder's grandson, who worked as a shorthand writer in the Houses of Parliament in the early nineteenth century, overlapping with Dickens himself in the early 1830s. This preface, written in 1824, proudly proclaims that "the excellence of the system has since been proved, in my own office, by a trial of many years, on a large scale, and I have not, in the course of that experience, found the necessity of any further alteration." It seems, then, that while he was learning shorthand in the late 1820s, Dickens would have been using one of the standard editions that were available in the early nineteenth century. If the price Dickens quotes is correct, ten and sixpence was a significant investment for him to make, considering that Pitman's Stenographic Sound-Hand, which came out only only ten years later in 1837, cost only four pence. Although the subtitle of Brachygraphy is "an easy and compendious system of shorthand," it was certainly not easy to learn.

Gurney sets out his symbols in three tables, which illustrate what the symbols stand for, and a set of rules, called "contracting rules," which tell the shorthand writer how to reduce the spelling of words before turning them into shorthand symbols. One of his tables, shown in figure 1, contains 24 symbols, called "arbitrary characters," in which a symbol stands for a particular words or set of words.

On this page the symbols represent what Gurney presumes to be keywords for legal/parliamentary stenographers, together with their derivational morphemes. In compiling the list, Gurney seems to have selected words which he considers to be frequent in the kind of texts he would expect shorthand writers to be writing—that is, legal and political speeches. He has also chosen words that are quite long, presumably to save the shorthand writer the trouble of spelling out the individual symbols of these long words. As can be seen in the left-hand column of the table, the symbol Gurney chooses to represent these words is the first letter of the word. Gurney's lack of attention to sound is evident in the use of the symbols "c" for "circumstance" and "s" for "statute" even though they both start with the same /s/ phoneme. The mnemonic principle is clearly to learn these forms by memorizing the first letter of the example word, not the sound.

Gurney then provides a second set of arbitrary characters, shown in figure 2, which he classifies as "Prepositions and Terminations &c." This second table is a bizarre mixture of non-lexical words with a mostly grammatical function, comprising auxiliary verbs ("had," "doth"), modal verbs ("may," "might") prepositions ("above," "below"), articles ("the"), pronouns ("I", "thou", "thee"), possessive adjectives ("thy"), a participle ("been") and a number of collocational strings such as "it is" and "it is not." It also includes symbols representing bound morphemes, such as the prefixes "ab-" and "ob-" and the suffixes "-able" and "-ible," as well as symbols for the affricate sound /ʧ/ and the fricative /ʃ/. Curiously, the list does not include the frequent pronouns "he," "she," "we," and "they" and a number of very frequent conjunctions like "as" or prepositions like "on." There are also individual symbols for five nouns ("plaintiff," "defendant," "foundation," "advance," "consequence"), an adjective ("covetous"), a verb ("magnify"), and four symbols representing the word "world" and three of its collocations. In this table a single symbol can stand for one or two morphemes, one or two words, or a mixture of morphemes and words. The first symbol at the top of the left hand column, for example, stands for two bound morphemes ("ab-" and "ob-"), and three lexemes ("observe," "observance," and "observation").

Fig 1.

Arbitrary Characters. (Gurney's Brachygraphy, 15th edition, 1825.)

The 75 symbols in the two Gurney tables are "arbitrary" in the sense that the symbols do not bear any relation to the words they are supposed to represent. They are, in fact, the equivalent of Chinese logograms—shapes that learners need to memorize and associate with the word they refer to—and they show that the Gurney system is at least partially logographic. The classification of logographic and phonographic writing systems⁶ is shown in figure 3.

Systems that represent languages (glottographic systems) can be logographic or phonographic or contain elements of both. Logographic writing systems like Chinese use logograms to represent word meaning, while phonographic systems like the Roman alphabet use letters to represent sound. The writing system of

Fig 2.

Prepositions and Terminations. (Gurney's Brachygraphy, 15th edition, 1825.)

Fig 3.

Gurney and Pitman stenography as glottographic writing systems. (Based on account of chapter 2 in Sampson, Writing Systems.)

Japanese uses elements of both (see Sampson, 208–32). Like Japanese writing, stenographies can also exhibit logographic or phonographic characteristics to different degrees. The Pitman system, which prides itself on not using arbitrary characters or logograms, is almost completely phonographic, whereas the Gurney system, which uses logograms representing words as well as letters representing sounds, has a foot in both logographic and phonographic camps. Dickens himself appears to have noticed the logographic nature of Gurney's arbitrary characters. He describes David Copperfield's attempts to read back his shorthand notes in the following terms:

The result of so much good practice was, that by and by I began to keep pace with Traddles pretty well, and should have been quite triumphant if I had had the least idea what my notes were about. But, as to reading them after I had got them, I might as well have copied the Chinese inscriptions of an immense collection of tea-chests, or the golden characters on all the great red and green bottles in the chemists' shops!

(DC, 505, ch. 38)

The reason why Dickens is comparing the characters to Chinese inscriptions is that they evoke the idea of being able to visualize the elements of a script but being unable to piece them together to sound out a word as a whole. Since the Gurney symbols have no link to sound, readers have no immediate access to the phonological code to help them understand what they represent and so they are helpless when trying to read them. By using the phrase "Chinese inscriptions," Dickens is lamenting Gurney's lack of phonography as a guide to comprehension.

The second main difference between the Gurney and Pitman systems is in their relationship to sound. The Gurney symbols that are not logograms connect to sound via letters, whereas the Pitman system posits a direct relationship between symbol and sound. Gurney's third table, shown in figure 4, shows 24 symbols standing for each letter of the alphabet and a particular word or words.

In the "Alphabet" column, each letter of the alphabet has its own symbol with the exception of four pairs of letters, which share a symbol. The letters < c > and < k > seem to have the same symbol for phonemic reasons—both letters can be pronounced using the /k/ phoneme. However, the letter < c > can also be pronounced /s/ as in the word "race." When transcribing words like "race," Gurney follows a phonemic principle and uses the symbol for the letter < s >. The < s > and < z > pairing is also problematic because they are not only different letters, but they also produce different sounds. They are both alveolar fricatives but /s/ is voiceless while /z/ is voiced. Giving these letters the same symbol means that the shorthand symbols are identical for words such as "rise" and "rice," which makes reading the symbols back much harder. The same applies to the pairing of the letters < v > and < w >, which have the same symbol even though, to modern ears, the consonant sounds that the letters represent are very different.⁷ The letters < i > and < j > also have the same symbol, which stands for < i > when it occurs at the beginning of a word, as shown in the transcription of the word "illustration" in the right-hand column; there is no illustration in the table for the use of the letter < j > but it does not seem to involve a phonemic principle because the symbol for < j > is not used to stand for the /ʤ/ phoneme at the end of the word "messenger" (see the letter < m > in the table).

Unlike the Gurney symbols, which stand for words or letters, the symbols of the Pitman system stand directly for sounds. In fact Pitman does not discuss letters at all, but follows John Walker's early attempt in Principals to define vowel and consonant sounds in terms of their manner of articulation. Pitman's system uses 24 symbols standing for single consonant sounds and 12 symbols standing for vowel and diphthong sounds. Pitman argues, correctly, that the fact that he has designed his system so that learners go straight from the symbol to the sound without mediation by letters is advantageous for them. Figure 5 shows this fundamental phonographic difference between the Gurney and Pitman systems by adding some detail to figure 3.

The bottom level of figure 5 shows that phonographic systems of writing are divided up into three types of script—syllabic, segmental, and featural. An understanding of these distinctions helps to explain why different types of shorthand are differently phonographic. Roman alphabetic writing falls into the segmental writing category, which uses individual graphs to represent individual sounds. In the case of English, the graph-sound correspondence is not as consistent as in other languages, with letters corresponding to a number of different sounds depending on the verbal context. Featural systems are systems which incorporate details of sound production into the graphic representation.⁸

Fig 4.

Alphabetical List (symbols, letters, words). (Gurney's Brachygraphy, 15th edition, 1825.)

Fig 5.

Gurney and Pitman stenography as logographic/phonographic systems. (Based on account of chapter 2 in Sampson, Writing Systems.)

Shorthand systems, like any writing system, also belong to different categories. Systems such as Pitman's, which include the detail of sound production in their graphic representation, are featural,⁹ while alphabetic systems such as Gurney's, in which there is a one-to-one relationship between a symbol and a letter, are segmental. This presents us with a schematic difference between the two shorthand systems. As a writing system, Gurney shorthand is closer to the Roman alphabet than it is to Pitman shorthand because its symbols stand for letters, not sounds; in the Gurney system, a symbol stands for a letter, and sounds are only revealed through the medium of letters. Overall, the fact that Gurney shorthand is partly segmental/phonographic and partly logographic, whereas Pitman shorthand is purely featural/phonographic and therefore much more transparent as a writing system, is crucial when we come to consider their respective contributions to the representation of orality.

A further crucial difference between the Gurney and Pitman systems is in the representation of vowels. Gurney only uses symbols for vowels in words whose initial letter is a vowel. Everywhere else, vowels are represented by the positioning of the consonant symbol, as shown by Gurney's abbreviating rules in figure 6.

Figure 6 tells us that in the Gurney system a visible character (a dot) will mark vowels only when they come at the end of a word. Vowels in a central position do not have an individual mark but are signalled by the positioning of the consonant following it. Inclusion or omission of symbols for central vowels seems to be at the discretion of the shorthand writer. In his preliminary instructions Gurney casually remarks that "it will be observed that in the spelling of words no particular regard is had either to the retaining or omitting of vowels." This somewhat loose treatment of vowels is confirmed by a number of examples in the practice materials which ignore the positioning rule and simply omit the vowel.

In the Pitman system, there are 12 symbols standing for vowel and diphthong sounds, so that vowel sounds are visible and the Pitman writer reads them directly off the page. This is not the case for the Gurney reader, who faces a set of consonant symbols and has to fill in the gaps between them with an appropriate vowel sound. Again, the different treatment of vowels in the respective systems obliges their users to treat the textual relationship between sounds and words in a completely different way.

Fig 6.

Gurney's rules for the representation of vowels.

Another extremely important difference between the Pitman and Gurney systems is in the type of graphic representation they use. Perhaps the most original feature of graphic representation of the Pitman system is that it makes use of the connection between the shape of a graphic mark and the way in which the sound is produced. Pitman describes his attempt at classifying vowels and diphthongs as being "according to the order of nature" (3). By this he means that he is categorizing sounds in terms of their physical articulation. So for Pitman the long vowels are classified first, beginning with the front vowel /ɪ/, which requires a small opening of the mouth, and moving on to the sounds /eɪ/ and /a:/, which involve a progressively wider opening of the mouth, and then on to the rounded back vowels /ɑu/, /əu/ and /u/.

A classification which relates shape to sound has two advantages. First, in relation to consonants, it enables Pitman to be economical with the number of graphs that he uses. For example, instead of using 18 different symbols for 18 voiced and voiceless consonants, he only uses nine, attributing them to voiced/ voiceless pairs that are articulated in the same way (/f/ and /v/; /k/ and /g/; /p/ and /b/; /s/ and /z/; /t/ and /d/; /ʧ/ and /ʤ/; /ʃ/ and /ʒ/; /θ/ and /ð/) and signalling the difference between a pairing by using a heavier stroke of the pen for the voiced consonant; for example, /f/ and /v/ have the same graph but with a heavier stroke for the /v/ graph because it is voiced. The second advantage is that the relationship between the way a graph is written and what it refers to is more direct. In the case of /f/ and /v/ there are two types of reinforcement—the voicing of the consonant corresponds to greater pressure on the pen (a tactile connection between the writer and the symbol) and a more visible mark (a visual connection between the writer and the symbol). Pitman exploited these synaesthetic relationships between the shape of a symbol and its sound wherever he could; for example, he reinforces the contrast between short and long vowels by giving short vowels a light dot or stroke and long vowels a heavy one. He gives four diphthongs a curved shape because "their wheel-about shape corresponds well to their turn-about sound" (3). Pitman makes this point about synaesthetic reinforcement even more explicitly to his readers by cleverly varying the typeface of his own text in order to illustrate that a shape can be made to look like what it sounds like (see Figure 7).

Fig 7.

Symbolic graphic variation in Pitman

Gurney, by contrast, frowns on the use of different weight of marking. He writes "the practice of placing a character above or below the line; and also of varying it, by thickness or length, I totally expunge; as it could never be written swiftly, with any tolerable degree of precision" (16). This statement epitomises the difference between the Gurney and Pitman approaches. Pitman claims that his use of graphic variation is more economical because it uses the same graph with different variations to indicate different sounds.

Having looked at the details of the Gurney and Pitman systems in terms of their phonographic differences, we now turn to the act of reporting itself and what the shorthand reporter was actually recording, particularly the "verbatim" reporting that critics claim to be a stenographic source of Dickens's orality. According to this argument, Dickens was listening carefully to and recording the sounds of the courtroom and this kind of reporting had a phonographic character that was at the heart of his "phonetic spelling" of fictional speech.

In reality, the kind of verbatim reporting that Gurney shorthand reporters were engaged in was a very limited aural experience and had very little to do with sound. Shorthand reporters certainly operated with what Monika Fludernik calls an "idealised requirement to fidelity" (16), but fidelity to what, exactly? For a law reporter being faithful to the utterance means reporting accurately the actual words of what was said in court¹⁰ or Parliament but not necessarily how it was said. Court reporters do not provide information about particular sounds in their reports. Put another way, verbatim stenographic transcription does not involve "orality" in the sense of nonverbal conversational detail. Quite the reverse in fact. Stenography is not like musical notation, which prioritizes sound production and represents ideal sounds with additional instructions for the production (speed, volume, style, and so on). Most stenographers are selective listeners, who are really only interested in the segmentation of the stream of speech into words. Scripts derived from stenography do not resemble the kind of transcript that aims to capture a high level of linguistic detail. One only needs to look at the kind of modern transcripts used to analyze courtroom discourse for research purposes to understand how much information verbatim stenography ignores. An example of just such a script, taken from Drew (486), and written in the Jeffersonian style (a style of description designed to record both the verbal content of a conversation and also non-verbal details regarding the manner of speaking):

6 Witness: I don't remember what he said to me that night

7         (1.2)

8 Counsel: Well yuh had some uh (p) (.) fairly lengthy conversations

9 with thu defendant uh: did'n you?

10         (0.7)

Here we see how Jeffersonian transcription aims to provide the discourse analyst with as much linguistic information as possible about a speech event. In this example we find fillers like "uh", hesitations marked by "(.)", significant pauses marked in tenths of a second such as "(1.2) and (0.7)", and volume signalled by underlining ("don't remember"). Gurney shorthand script would not codify any of these elements because whoever commissioned the law report would be interested in what was said, not how it was said.¹¹ This is confirmed by the only surviving transcript of a Dickens shorthand report¹²—the Jarman v Bagster judgment—which only includes the words of the judge written in standard English with standard spelling and no markers of paralinguistic information.

Far from making its shorthand learners more aware of orality, the alphabetic Gurney system may have done the reverse, training learners to ignore details of spoken interaction that they considered irrelevant for the purposes of the final transcript. Since the practice of courtroom stenography actively encouraged Dickens not to listen for oral conversational detail, it seems unlikely that stenography made Dickens more proficient in representing the oral qualities of words as they are pronounced, as critics have claimed. If anything, it did the opposite, making him proficient in discarding paralinguistic features that were irrelevant to his final transcript.

Nor can it be argued that Dickens was imitating the different idiolects of the Doctors' Commons courtroom. The cases heard at Doctors' Commons certainly involved evidence from a range of witnesses from all echelons of Victorian society—baronets, heiresses, members of the armed forces, housekeepers, prostitutes, churchwardens, servants, a bailiff, a clergyman, a surgeon, a clerk, and a stableowner¹³—but their actual voices were never heard giving evidence in the courtroom because ecclesiastical court procedure required plaintiffs and defendants to provide testimony via written depositions only.¹⁴ Dickens only heard the evidence of ordinary people through the mouths of lawyers as they read out witness statements, which themselves only reported the words of witnesses in indirect speech. What is more, as the Jarman v Bagster text confirms, the direct speech that Dickens eventually wrote down in his transcript was what the judge said at the end of a case—a judgment that included a summary of what the lawyers had been arguing. So in his ecclesiastical reports Dickens was essentially transcribing prepackaged testimony that had already been written in a deposition, voiced by lawyers and then filtered through the mouth of a judge. If ever the words of witnesses came to be transcribed by Dickens at Doctors' Commons, they arrived as what might be termed "third-hand orality," embedded in other people's texts and transformed by other people's voices. This is a long way from the spontaneous conversational language of the Pickwickians.

Putting the arguments of the first two sections together, there are many reasons to doubt that Dickens's knack for "infusing writing with all the immediacy of the moment of oral utterance" (Kreilkamp, Voice 77) developed as a consequence of what he was writing in the courtroom. Even if he had wanted to register the manner in which people spoke, the Gurney system had no graphic means of doing it. Half of the symbols that he had memorized (arbitrary characters) were logograms that stood directly for words, not sounds, and the remaining symbols he was using (alphabetical characters) stood for letters, not sounds. Furthermore, he was not writing down symbols to represent vowel sounds because the Gurney system did not use them. Finally, he was a word recorder in the courtroom, not a voice recorder, and he was only recording the words of the judge in his final judgment, not the oral detail of courtroom discourse in general and certainly not the lively direct speech of ordinary people. Overall, if we are looking for a source for a hypothetical connection between stenography and orality, we will not find it in a presumed phonographic character of the shorthand he had learned or the symbols he was using.

My argument in this article is that the key to understanding the connection between Gurney shorthand and Dickensian orality is not in what he wrote as a stenographer or how he wrote it but in the way it forced him to think stenographically about his writing. In other words, as Dickens navigated his way through the "sea of perplexity," the impact of learning and practicing shorthand on his writing habits created a "Gurney mindset" because of the unusual way it trained him to process language. When moving from hearing sounds in court to representing them graphically in shorthand, the difficulty for the writer was not understanding the words he was listening to but mentally processing them in the way that the Gurney system required. As we have seen, this processing involved shortening words through the omission of vowels and then trying to read back the abbreviated consonant symbols.

Although a great deal of the Gurney manual is taken up by examples of how speech has to be reduced, there are no actual rules concerning how the writer is supposed to do the reducing. Gurney's 11 "abbreviating rules" actually have very little to do with abbreviation, for there are no specific instructions for word reduction in the "abbreviating rules" section. Gurney provides only a vague statement of intent—"the other column on the first page is designed to give the learner an idea of the manner of joining the letters and of spelling words: but need not be committed to memory" (18). The learner is therefore supposed simply to look at the central "Letters" column and infer how reduction ought to take place. Table 1 is an example taken from the practice materials of the Gurney manual showing how a legal text would be turned into symbols and what these symbols "spelt" in Roman letters.

The text of "spelt symbols" on the left shows how the symbols would be spelt out in alphabetical letters and the text on the right shows what the end product of their reconstruction would have been; the letters in bold are the letters that need to be added to the Gurney original. In order to expand the text of "spelt symbols" into a full text the writer has to make guesses as to what the reduced words can be expanded into, rather like decoding the modern-day text message. These expansions could involve making the following types of change:

Insertions

• • word-initial vowels, e.g., from < ny > to < any >

• • short central vowels, e.g., from < cnsts > to < consists >

• • long central vowels, e.g., from < kping > < keeping >

• • long central dipthongs, e.g., from < oblging > to < obliging >

• • word final vowels, e.g., from < stplat > to < stipulate >

• • the aspirated < h >, e.g., < as > to < has >

• • redoubling of a consonant, e.g., from < grs > to < gross >

Changes of consonant letter

• • consonant letters standing for affricates, e.g., from < plejs > to < pledges >; from < csn > to < caution >

• • consonant letters standing for sibilants, e.g., from < ofns > to < offence > (same sound) or < gs > to < gaze > (different sound)

Table 1.

Gurney "spelt symbols" and full text.

Some words are rather difficult to predict from their consonant skeleton. For example, the consonant sequence < ths > could initially be turned into < this >, < those > or < these >. The reader sees that the subsequent word "persons" is plural so the choice would then be between < these > or < those > and the final decision would rest on whether < these > or < those > best fitted the sense of the sentence as a whole. Overall, the predictability of < ths > rests on a set of lexical, syntactic, or semantic choices inferred amost simultaneously from the context of the initial word.

There are therefore many potential variations for shorthand writers to keep in mind when reading back shorthand symbols and writing out the words they are supposed to represent. Table 1 shows that most of the expansion of the consonant skeleton involves the insertion of vowels and diphthongs. For example, if Dickens as a reporter had heard the word "full" spoken by the judge, he would have written down the symbols standing for the consonants < –,>, as in the sequence < and to gv fl asrns >. When reading back his shorthand notes after the hearing, he would have to expand the two consonants into a complete word by carrying out a mental check of the possible vowel or diphthong sounds that could fit the gap between the symbol for < f > and the symbol for < l >. He would do this by looking at the position of the second consonant symbol relative to the first and running through the options ("fall," "fail" or "fell" in the upper position; "fill" or "file" for the middle position; "foal," "full," "foul," "foal" or "folly" for the lower position) and choosing the most appropriate word to fit the context. In this mental search for the right vowel he would be thinking about possible vowel sounds in relation to possible words. This testing of vowel sounds is a restricted process, involving a limited number of sounds that can only be used in a limited number of contexts, the analysis of which is known as phonotactics.¹⁵ The habit of reading back Gurney shorthand would have turned Dickens, and any user of the Gurney system, into a formidable phonotactician.

There are a lot of comparisons that can be made to explain the process of writing Gurney shorthand and reading it back. The conveying of information through writing consonant symbols is comparable to the writing of text messages, albeit with an unfamiliar script. Reading back Gurney shorthand is like playing a game of Scrabble where there are only consonants on the board and the player only has vowels on his or her rack; this involves a sophisticated sense of phonotactics. However, perhaps the best analogy for Gurney processing is Douglas-Fairhurst's metaphor of the "concertina" (62). The concertina represents a mental process, and movement of the concertina metaphorically captures what happens to words when reading and writing Gurney shorthand. When writing shorthand in court, the Gurney stenographer mentally compresses the concertina to remove the vowels from a word, leaving a consonant skeleton of shorthand symbols in his notebook. When transcribing the skeleton of symbols into longhand, he mentally expands the concertina, and inserts vowels and other consonants into the skeleton to produce possible words. While he was transcribing law reports day after day, Dickens was honing his ability to encode and decode words using the concertina and by the time he came to compose his own fiction, these processes had become a fixed part of his mental orchestration of words.

We also need to consider the context of reading back shorthand. Stenography was a solitary working practice. Although Gurney stenographers were able to read each other's shorthand, the speed with which reporters had to work in order to meet deadlines meant that each stenographer only read his own work, and Dickens would have spent many hours alone poring over his own symbols and reconstructing what he had written. This would have involved verbalization of what he was reading. Vowel-testing for Dickens was a reading process in which he was sounding out the vowels to himself. The possible combinations of consonant and vowel sounds running through the Gurney stenographer's head during readback were his/her own idealized representation of sounds (phonemes), not genuine phonetic realizations of sounds (phones).¹⁶ Dickens sounded out vowels through "inner speech"¹⁷ by articulating the phonemes of possible words to himself subvocally. The sounds he was transcribing during this abstract, monologic, and highly personal process had very little connection to the actual pronunciation of ordinary words by ordinary people in the real world.

The contextual and cognitive solitude of his shorthand training would have reinforced Dickens's personal identification with the Gurney system. In a much quoted passage Dickens describes how, when listening to a particularly boring speech, he would "sometimes beguile the tedium of the moment by mentally following the speaker in the old way; and sometimes, if you can believe me, I even find my hand going on the table-cloth, taking an imaginary note of it all" (Speeches 347–48). This anecdote is illustrating more than just a Pavlovian reflex. On the one hand, it shows how the Gurney mechanism had provided Dickens with an automatic connection between mind ("mentally following the speaker") and imagination ("an imaginary note"), while on the other it shows that the isolation and intensity of working with the Gurney system had given him an instinctive, embodied attachment to shorthand even when he did not need to use it. Leah Price speculates that "the days and nights that Dickens spent studying an 1824 reprint of a 1750 manual must have felt doubly galling thanks to the publication, in 1837, of Isaac Pitman's new method, Stenographic Soundhand" (43). However, Dickens seems to have been quite indifferent to the post-1837 rise of Pitman. He continued to see the Gurney system as a route to professional advancement and, remarkably, even found the time to teach it to Arthur Stone in 1859—"It is highly desirable—above all things—that you should now get to the Short Hand. If you can begin with me here, at 10 tomorrow morning, do" (Letters 300). Dickens's emotional attachment to the clumsy Gurney system resembles our own attachment to the notorious inefficiency of English spelling. There are times when the quirkiness of a system, our familiarity with it, and our enjoyment of it will trump its lack of logic.¹⁸ Just as influential English speakers who have mastered the intricasies of English orthography are unlikely to undermine a system to whose mastery they owe their influence, Dickens may have detected the cognitive advantages that the Gurney system had given him and wanted to pass them on.

What evidence is there of the Gurney concertina being played in Dickens's fictional speech? Looking at Pickwick we see the Gurney rules of vowel abbreviation partially at work in Sam Weller's elimination of vowel sounds in words like "gen'l'm'n," "cert'nly," "reg'ler," "singl'r" and in the elimination of the penultimate letter < e > in the shift from "creatures" to "creeturs." The shortening of the < ng > to < n > as in words like "startin'" or "lodgin's" is also stenographic since Gurney uses a single arbitrary character for < ng >. There are also Gurney expansion mechanisms involved in the representation of Sam's speech. The nonstandard spelling of the vowels in "wos" (for "was") or "buzzum" (for "bosom") manipulates the central vowel in the manner of a Gurney reader testing hypotheses for possible words. The word "buzzum" also shows an alternative consonant (< z > instead of < s >). This is also Gurney-like since the letters < s > and < z > have the same Gurney symbol, and so the shorthand reader might be prompted to try out both letters when testing for possible words.¹⁹ The same applies to Sam's well-known inversion of the < v > and < w > letters ("wery" for "very"); there is only one Gurney symbol for both the letters < w > and < v > and this binary choice may have facilitated their interchangeability in Sam's speech. There is also a stenographic influence in the nonstandard insertion of the letter < r > in some of Sam's spoken words. In the Gurney system the symbol for < r > has to represent the letter < r > even when the < r > is not pronounced, like the final < r > in the word < messenger >. So the use of < r > in the spelling of a word like "formiliar" seems to be an invention derived from Gurney principles—Dickens is including an < r > letter in a position where the /r/ sound is not pronounced but where < r > might easily be used in the spelling of a similar sounding word such as < fort >. The addition of consonants like the extra < d > in "widder" (standing for "widow") is also stenographic because in the Gurney system double consonants are reduced to just one symbol, and so the same two Gurney symbols would represent both "wid" and "widd."

However, it also needs to be recognized that just as there are Gurneyesque influences in the spelling of some areas of Sam's speech, there are just as many areas in which Dickens's non-standard spellings do not follow Gurney stenographic principles. These include Sam Weller's maintenance of the vowel in the onset syllable of "gen'l'm'n," "cert'nly," "reg'ler," "singl'r" and the elimination of consonants like the < t > in "mas'r" or the "v" in "ha'nt" (a Gurney stenographer would not have omitted these consonant symbols); changes in consonants like the shift from < f > to < r > in "arter" or the < n > to < bl > in "chimbley" (these changes would not be permitted by Gurney); and the addition of consonants like the < h > in "hemperor" or the spelling of the central vowel < ee > in "creetur," which would have been eliminated in Gurney shorthand.

Another interesting example of Gurneyesque speech in Pickwick is that of Smorltork, the Russian count. Like the representation of the Wellers' Cockney, Dickens's use of phonetic speech is designed to mimic a Russian accent.²⁰ Like Sam Weller's "formiliar," the character name "Smorltork" shows the same use of the < r > symbol in a word (small, talk) where it is not needed, and there are a number of examples in the count's speech where he appears to be verbalising the operations of a Gurney shorthand writer²¹:

"They are here", added the count, tapping his forehead significantly. Large book at home—full of notes—music, pictures, science. Potry, poltics, all tings". … And down went Mr.Pickwick's remark, in Count Smorltork's tablets, with such variations and additions as the count's exuberant fancy suggested, or his imperfect knowledge of the language occasioned.

(560; ch. 30)

Here Smorltork is behaving like a Gurney shorthand writer, but with a difference. He is taking notes of what Pickwick is saying but he is also verbalizing them as if they were Gurney's "symbols as spelt." Each of the unusually spelt words partially follows Gurney abbreviating rules. For the word "poltics," he eliminates the central vowel letter < i >, for "things" he turns the < th > into a single letter < t >, and for "poetry" he reduce the triphthong /əuə/ to the single letter < o >. There is more to come:

'What you say, Mrs. Hunt?' inquired the count, smiling graciously on the gratified Mrs. Leo Hunter, 'Pig Vig or Big Vig—what you call—lawyer—eh? I see—that is it. Big Vig'-- and the count was proceeding to enter Mr. Pickwick in his tablets, as a gentleman of the long robe, who derived his name from the profession to which he belonged, when Mrs. Leo Hunter interposed.

'No, no, count,' said the lady, 'Pick-wick.'

'Ah, ah, I see,' replied the count. 'Peek—christian name; Weeks—surname; good, ver good. Peek Weeks. How you do, Weeks?'

(Bevis 205)

This passage shows Smorltork testing out possible words in the manner of a Gurney stenographer. His change of the central vowel letter < i > to < ee > in "Peek" and "Weeks," as well as the addition of the final < s > in "Weeks," would all be hypotheses that the Gurney shorthand writer would produce when imagining the words that could spring from < pkwk >—the Gurney spelling of the word "Pickwick."

The effects of the Gurney mechanism are not confined to character speech. Since the mechanism develops a general cognitive ability for letter and sound manipulation, it facilitates the use of poetic devices that critics have long associated with Dickens. Dickens uses the memorable phrase "shirking and sharking" (BH 53) to describe the laziness and dishonesty of the legal profession. This is an example of consonance in which one can also hear the Gurney stenographer alternating two vowel sounds inside the consonant skeleton of /ʃ/ and /k/, heightening the aggressive impact of the consonants. Likewise when Dickens describes Master Adolphus in The Haunted Man as going "to and fro, in his little oilskin cap and cape," we are not only presented with the rhythmical change of sound between the long vowel in "to" and the diphthong in "fro" (with both words spelt with a final < o >) but also the change in vowel sound between the consonant sounds /k/ and /p/ in "cap" and "cape." The phrase "cap and cape" represents the sounding out of the shorthand spelling of a word spelt < kp > in Gurney shorthand.

So we find traces of the Gurney mechanism in many areas of Dickensian style, not just dialogue and conversation, and they affect the reader in different ways. In relation to speech, the non-standard orthography forces the reader to do some kind of mental guesswork about a strangely spelt word in order to work out what word it is supposed to represent, while in relation to narrative he juxtaposes familiar words with familiar spellings and invites the reader to enjoy the effect of the manipulation of the sound. Overall, the Gurney-driven stylistic effects that Dickens brings to bear in character speech and narrative show that he is getting his readers to do what the Gurney manual forced him to do as a stenographer, namely to read both speech and narrative differently, playfully, and creatively.²²

In conclusion, the legacy of Gurney stenography for Dickens was not graphic mimicry of voices or sounds but speed of writing, accurate reporting of words in direct speech, and a distinctive cognitive predisposition which allowed his phonotactic awareness to flourish. If, as Steven Marcus suggests, the practice of stenography enabled Dickens to see a script as a different system because he could inscribe the same concepts using two different forms of writing (the Roman alphabet and Gurney symbols), the learning of the Gurney system enabled him to construct a script differently and with a creative architecture that could not have been achieved if he had learned the Pitman system. This construction is both piecemeal, extending to narrative areas of Dickensian fiction as well as its lively orality, and multifunctional, at times mimetic of Gurney spelt symbols and at other times visible by inference in the use of figures of speech such as consonance and assonance. The fact that these alphabetical/stenographic effects are distributed as an uneven patchwork across different genres suggests that future stylistic research needs to look extremely carefully at exactly how and where the Gurney mechanism operates in Dickensian fiction.

The historical implications of the Gurney mechanism are equally important. If the unique design flaws of the alphabetical Gurney manual are responsible for the cognitive mechanism that links stenography and Dickensian orality, it is difficult to see how Pickwick "inaugurates the phonographic history of Victorian fiction" (Kreilkamp Voice 77). When looking for neat historical patterns, it may be comforting to see the Pickwickian verbal explosion as somehow sparking off the phonographic turn of the Victorian era, yet the orality of Pickwick could equally be viewed as completely discontinuous with it—as the product of the unique and innovative alphabetical construction of literary oral effects brought about by the creative learning of an unusually difficult alphabetical system. A realistic critical assessement of the position of Pickwick in relation to the diachrony of the orality/literacy interface requires a more nuanced approach that takes account of the details of Brachygraphy's unique logographic/alphabetical features and abbreviating systems and how Dickens learned and used them.