Different Times, Different Materials and Different Purposes: Writing on objects at the Grand Arcade site in Cambridge

Like many other topics the subject of the materiality of writing in 18th–20th century Britain has received relatively little attention, principally because it has been conceived of as part of an unproblematic “familiar past” (Tarlow and West 1999) that is perceived as similar to the present or sufficiently well understood through other sources that archaeology does not have a significant contribution to make. There are also other major issues that differentiate the 18th–20th centuries from earlier periods, most notably that the material culture is predominantly mass-produced in a way that few earlier examples of writing are; additionally many types of material culture of the period are truly global in extent. In contrast with earlier periods in Britain where there is relatively limited archaeological evidence for writing, the problem here is that the amount of data is often too large. One way to approach this richness of data is to eschew the more broad brush quasi ‘culture-historical’ approaches often adopted for earlier periods, where material spanning several centuries and large geographical areas is studied in order to generate a large enough ‘corpus’ of material to make meaningful comments. Instead the evidence from 18th–20th century Britain allows us to work on a much more intimate scale of individual households at particular points in time as represented by ‘feature groups’. Any such attempt to consider 18th–20th century writing must recognise that the dominant material upon which writing was produced was paper, which is rarely preserved archaeologically even in 18th–20th century contexts, although there are exceptions (Crook and Murray 2006:


Introduction
Like many other topics the subject of the materiality of writing in 18 th -20 th century Britain has received relatively little attention, principally because it has been conceived of as part of an unproblematic "familiar past" (Tarlow and West 1999) that is perceived as similar to the present or sufficiently well understood through other sources that archaeology does not have a significant contribution to make. There are also other major issues that differentiate the 18 th -20 th centuries from earlier periods, most notably that the material culture is predominantly mass-produced in a way that few earlier examples of writing are; additionally many types of material culture of the period are truly global in extent. In contrast with earlier periods in Britain where there is relatively limited archaeological evidence for writing, the problem here is that the amount of data is often too large. One way to approach this richness of data is to eschew the more broad brush quasi 'culture-historical' approaches often adopted for earlier periods, where material spanning several centuries and large geographical areas is studied in order to generate a large enough 'corpus' of material to make meaningful comments. Instead the evidence from 18 th -20 th century Britain allows us to work on a much more intimate scale of individual households at particular points in time as represented by 'feature groups' .
Any such attempt to consider 18 th -20 th century writing must recognise that the dominant material upon which writing was produced was paper, which is rarely preserved archaeologically even in 18 th -20 th century contexts, although there are exceptions (Crook and Murray 2006: How to cite this book chapter: Cessford, C. 2013 66-69, 80-85;Dickens 2001: 117, 124). This is true of many periods, where post-depositional processes and environmental conditions have often largely destroyed all traces of the dominant writing material. In the case of the 18 th -20 th centuries much of the material has been recovered from material dumped in below ground features where conditions have destroyed all paper, although in contrast to earlier periods we have a much better understanding of 18 th -20 th century writing on paper since vast corpora are preserved in libraries and archives. Despite the poor preservation of paper in 18 th -20 th century archaeological contexts, many forms of writing that do survive archaeologically on more durable mediums are related in some way to the dominant paper medium. That the relationship between writing on paper and other mediums is often complex and ambiguous is perhaps best illustrated by an example found on a stone at the Nine Ladies stone circle in Derbyshire, where a graffito of the name "Bill Stumps" is incised onto an outlying orthostat of a Bronze Age stone circle (Figure 1). The incising of this name mirrors a fictional incident in Charles Dickens's The Posthumous Papers of the Pickwick Club published in 1836-1837, raising the question of whether the writing on the stone was inspired by Dickens writing on paper or whether Dickens was inspired by the inscription. Dickens does not appear to have visited the area until after he wrote the book so it is concluded that his writing published and circulated in a paper-based form inspired the graffito (Guilbert 2001). Although paper rarely survives archaeologically in 18 th -20 th century contexts, it has recently begun to be viewed from a more archaeological, or at least material culture standpoint. Particular attention is being paid to the materiality of paper-based writing, of the 19 th century (Hack 2005;Hall 2000;Marsden 2006;von Mucke 1999), as part of a general 'material turn' (Pykett 2005) in Victorian studies inspired largely by the work of Asa Briggs (1988) who asked historians to contemplate Victorian materialities, not least because the Victorians themselves were fascinated with objects and things.
While English is, unsurprisingly, the dominant language encountered in inscribed material culture from British sites of this period, a range of other European languages is occasionally attested, particularly French, but the next most common language is Chinese (see below). The majority of, but not all, examples of written objects during this period were mass produced. In theory this resulted in the production of virtually identical examples, and this has implications for many of the themes relevant to the materiality of writing. Prior to the 18 th century many, perhaps most, examples of writing that individuals encountered were effectively unique. This constitutes a markedly different type of encounter and it is notable that the earlier types of writing that were mass produced, most obviously coins, are largely absent from considerations of the materiality of writing. 18 th -20 th century mass production is, however, counteracted to a certain extent by the fact that some of the 'mass production' was relatively small-scale with localised distributions where products travelled at most a few dozen miles, whilst some types are truly global in their reach with examples found distributed around the world (see below). Additionally the great expansion of choice in some types of material at this time, such as ceramics, meant that, although produced on a massive scale, they might well be locally unique and restricted to a single household in an area. Such considerations must underpin the nature of the particular engagements with writing that will be presented subsequently.

Feature Groups
Archaeologically, material remains of writing are with a few exceptions relatively rare in Britain prior to the 18 th century. As a result, in bringing together enough material to enable meaningful comment most considerations of topics such as this generally have a relatively broad temporal and geographical scope. These tend to be 'culture-historical' in their approach, emphasising similarities and broad patterns (e.g. Evans 1987;Okasha 1995). A rather different approach is possible for 18 th -20 th century Britain, and indeed much of the rest of the world. Around the middle of the 18 th century a significant change occurs in the nature of the archaeological record in Britain. Increasing numbers of short-term deliberate depositional events survive, frequently containing hundreds of 'items' that can broadly be interpreted as 'feature groups' , closed assemblages of domestic artefacts discarded as a single deposit (Barker and Majewski 2006: 207;Fryer and Shelley 1998;Pearce  2000). These are interpreted as 'household clearance' events, such as those described in 19 th -century fiction in which they are characterised as profoundly brutal and disturbing (Trotter 2008). Objects that had been viewed by the Victorians as 'household gods' (Cohen 2006) imbued with personal meaning and social memory became simply commodities with an exchange value, and in the case of the material in the archaeological record simply 'stuff ' that is just waste matter (Trotter 2008) except in the contexts of deposition and potentially subsequent archaeological recovery. The phenomenon of feature groups presents investigators studying writing from a material cultural perspective with various interpretive possibilities and challenges. The contextual richness of such deposits means that they can become the primary analytical unit, rather than more spatially and temporally diffuse entities such as sites or cultures. These feature groups lend themselves to consideration through a form of "thick description", which does not look at material in isolation but takes account of context so that the things become more meaningful to an outsider (Geertz 1973), and the preservation of detailed archaeological associations provides a wealth of information about the meaning(s) of objects when situated in their various contexts.
Deposits of this type are attested in earlier periods in Britain but prior to the 16 th century they do not occur with any frequency. Only in the mid-18 th century do they become a more common occurrence. The increase in levels of discard in 'feature groups' at this time is probably linked to a consumer revolution, where in contrast to earlier periods dominated by scarcity and frugality there was a marked increase in consumption of a wide range goods and products by individuals from different social and economic backgrounds (Bermingham and Brewer 1995;Brewer and Porter 1993;Fairchilds 1993). This consumer revolution was fuelled by competitive emulation whereby individuals and groups lower down the social scale sought to imitate those higher up (McKendrick et al. 1982) or the restructuring of social relations particularly with regard to the changing nature of the bourgeoisie who owned the means of capitalist production, and to a growing and more assertive middle class (McCracken 1990). However, discussions of such assemblages often employ, albeit implicitly, the "Pompeii Premise" (Binford 1981;Schiffer 1985), assuming that the deposit represents a single moment frozen in time and that whoever made the deposit was also the original owner and/or user of the material. Once quantified these assemblages are used as the raw material for discussions of a host of themes, including social status and gender relations. Minimal consideration is usually given to the fact that the dumped material has probably been carefully selected. Still fashionable or valuable material was probably saved for further use, either on the same site or elsewhere (Johnson 1996: 182-183). This becomes apparent when the material from large assemblages is compared to that derived from other types of context, such as those related to middening and night soiling. The relative proportions of different material types and wares in the different types of deposit vary markedly, demonstrating that discard in 'feature groups' was carefully organised.
The nature of the assemblages, where objects are often complete or substantially complete and where broken typically consist of large unabraded fragments that can readily be refitted, means that material from these features can be quantified in terms of a count of 'minimum number of items' (MNI). This method is relatively straightforward, although not entirely unproblematic, for certain types of material such as ceramics (Brooks 2005), glass (Willmott 2002), clay pipes and worked bone objects. Based upon these MNI counts and the number of items with writing on them it is possible to calculate the percentage of writing-bearing items from the overall assemblage ( Table 1). There are a number of complicating factors such as incomplete items and items that bear more than one form of writing, which affect the quantification. The quantification used in this study specifically excludes categories of material that never or rarely bear any writing and focus principally upon ceramics, vessel glass and clay tobacco pipes.
The general availability of well surveyed cartographic evidence from the mid-18 th century onwards means that such assemblages can almost always be linked to particular plots or properties, while documentary sources mean that in many, but by no means all, instances land can be linked to known households. While there are occasional instances where some or all of the material deposited may originate from outside the particular plot in which it was recovered, the composition of most of the assemblages strongly indicates that they relate to a single household and originated on the plot where they were recovered. Such 'household archaeology' which developed in the 1970s (Wilk and Rathje 1982) has been extensively applied to archaeological remains of the post-1800 period (Allison 2003;Barile and Brandon 2004;Beaudry 1999;King 2006). The 'household' commonly consists of a nuclear family plus other elements, such as extended family members, household servants, employees and lodgers but can also encompass larger entities such as large businesses that included dozens of staff members. Nonetheless, it is clear from contemporary documentary sources that such large business 'households' were still viewed in familial and paternalistic terms (Roberts 1979). These feature groups can therefore be understood as representing assemblages of artefacts that relate to a single household and were discarded at a particular point in time, although the individual artefacts and assemblages also possess longer term 'biographies' (see below), and as such are an admittedly biased sample of the material culture of a given household.
By undertaking a detailed study of inscribed objects in a series of feature groups in what follows below, it is possible to arrive at a more nuanced understanding of writing materials and associated practices, in contrast to accounts produced by considering written evidence from a much wider geographical and temporal distribution. The material in this chapter derives from the large-scale excavations covering 1.5 hectares undertaken at the Grand Arcade site in Cambridge (Figure  2), by the Cambridge Archaeological Unit in 2005-2006 (Cessford andDickens, in preparation). This site is located on the edge of historic Cambridge, lying mainly in a suburb outside the town boundary known as the King's Ditch. The assemblages discussed here relate to a single 'street block' or group of plots bounded by street lines (Conzen 1960: 5), bounded by St. Andrew's Street, Downing Street, and the King's Ditch and its successor St. Tibb's Row.
While the writing on objects can be categorised in many different ways, two particular distinctions appear to be particularly significant. These are: • Writing that is primarily visual or primarily tactile • Writing that was apparent during normal usage of the item and writing that was concealed during normal usage of the item Of more than 40 feature groups investigated eight are considered here, although two of the feature groups are related, so this study effectively comprises six groups (Table 1). The features have been selected to provide a chronological range covering the longest possible period and also to include those with the more informative examples of writing.

Francis Tunwell's Planting Bed, F.6425
The first feature group to be considered is a planting bed dug in a large garden c.1760-1790, probably when it was being leased by a local merchant Francis Tunwell (Figure 3). A quantity of glass and pottery vessels and fragments was deliberately deposited in the base of the planting bed to act as a 'percolation fill' (Cotter et al. 1992: 161, 307-309, 450) to aid drainage. Among the glass vessels, only one was marked, having a rounded oval bottle seal with the text "PYRMONT WATER" around a crowned shield with the coat of arms of the principality of Waldeck-Pyrmont in Germany. Pyrmont mineral water became popular in Britain in the early 18 th century and by 1730 over 70,000 bottles a year were being imported into London (Hembry 1990: 176). The use of the English term "WATER" rather than the German "WASSER" indicates that they were produced primarily for export rather than local consumption in Germany, and they are relatively common finds in both Britain and North America in the period c. 1720-1770(Noël-Hume 1971. Mineral water from different springs supposedly had their own distinctive medicinal properties and consumers selected the water that matched their needs. To guarantee its authenticity Pyrmont water was exported in distinctively-shaped bottles with seals embossed with a crest and the name of the water put on at the time of manufacture. This particular bottle seal is poorly executed, being   badly aligned and the word "PYRMONT" is almost completely indecipherable. Related to an inherent problem with the technology of glass seals, illegibility and misalignment was not uncommon. Seals begin to be applied to glass bottles around 1650. A warm blob of glass would be applied to the body of the bottle and then impressed with an engraved metal die. As the body of the bottle was rounded there were often problems with the edges of the seal not being fully impressed, as seen in our case. Similarly, if the die was applied whilst the glass blob was too runny the impression would be blurred. If the glass was too solid the impression would also be unclear. Given that sealed bottles required the creation of a metal die and the seals were rather time-consuming to apply, such bottles were considerably more expensive than unmarked examples and were always a small minority of those produced and in use, and this is reflected in their archaeological frequency in assemblages. Accounts from 1676 indicate that sealed bottles cost 4½d apiece, whereas plain examples were only 3½d (Thorpe 1938). By the time this particular bottle was exported to Cambridge, Pyrmont Water was a well-established international brand; the writing was only one element in a package of distinctive bottle shape, coat of arms and text, so it was not necessary for the writing to be legible. Although technically tactile the glass seal must be viewed as visual writing, as it would not have been touched during normal use. The seals were placed low down on the body or high up on the shoulder of the bottle in locations that would have been awkward to hold the bottle by; additionally the writing and designs on seals were not sufficiently large or pronounced enough for their detail to be 'read' tactilely. Indeed their positioning appears deliberate to ensure that they did not impinge upon the grip of the individual holding the bottle, ensuring that they remained visible. The seals are not legible at distance and it is likely that only the pourer could have read them. Also found in the Francis Tunwell deposit were six clay tobacco pipe stems. These bore the relatively ornate 'Wyer' style decoration (Walker and Wells 1979) and the name and location of the manufacture, "S.WILK / INSON, / Camb. " or "S.WILK-/ INSON, / Cambg. ". Samuel Wilkinson was active in Cambridge from at least 1762 until his death in 1787. His practice of marking his pipe-stems was an innovation locally, as most 18 th -century Cambridge pipemakers did not mark their pipes or only used simple initials on the side of the spur, a small projection at the base of the bowl. Wilkinson's stem marks were much more visible than spur marks, but are frequently quite poorly executed. Given this and the fact that the mark was placed 80-100 mm from the bowl in the area where the smoker would commonly grip the stem, engagement with writing here would have been primarily a tactile experience. The relatively poor execution of the words was a result of the fact that they were applied by roller stamps which produced variable results. Nevertheless, the mark was arranged so that the text could be read by the smoker holding the pipe, so the visual aspect was still clearly of some importance. This visual importance is also confirmed by a few pipes from other contexts that do not bear Wilkinson's name but some other slogan, such as one marked "PARKER / for ever, / Huzzah", probably produced during the parliamentary election campaign of William Parker Hammond in 1783. It is impossible to be certain why Wilkinson marked his pipes when most of his local competitors did not, but the most likely reason is that they constituted a form of branding for a finer product, since Wilkinson's pipes were of better quality and finish than other contemporary local pipes. Wilkinson's stem mark is a form of branding similar to Pyrmont Water, but much more localised. From production to distribution, these particular pipes had travelled only about 200m, whereas the Pyrmont Water had travelled over 600 kilometres to reach Cambridge. The distribution of Wilkinson's pipes appears to extend not more than 25 kilometres from Cambridge. Wilkinson's marks were also more short-lived, spanning 25-35 years, whereas Pyrmont Water was common in Britain for almost a century. The longer temporal span and greater geographical spread meant that there was likely to be much greater brand awareness, recall and recognition of the Pyrmont Water. This is significant in terms of writing as material practice as it meant that the text of the Pyrmont Water could be recognised and in a sense 'read' without being legible.
Although this is the earliest feature group to contain examples of writing it was already present on two markedly different types of artefact that were used by the same household. The different types of writing function in markedly different ways, which were linked to the physical and technological nature of the items and to the different spatial and temporal spheres that the products operated within.

The Cock Inn Cellar, F.3029
The next group comes from the backfilling of a cellar in c.1828-1845, when an inn on this site was owned by John Purchas (1788-1848) and the proprietor was John Pike or William Bacon (Figure 4). The majority of the writing on ceramics relates to vessels marked "R Hopkins" on their underside (MNI 10), identifiable as Richard Hopkins, the cook for Gonville and Caius College (1805-1810) and Trinity Hall (1810). On some vessels of this kind this name is hand-painted (MNI 5) while on others it is transfer-printed (MNI 5), but all have similar moulding and blue hand-painted feather edge decoration. These vessels also have the impressed mark "TURNER", indicating they were manufactured by Turner's of Longton, Stoke on Trent (Hillier 1965).
Transfer-printing was invented in the mid-1750s and involved engraving a flat copper plate with the desired pattern; the plate was inked and pressed or transferred to a fine sheet of tissue paper. This was then applied to the pottery, which was fired at a low temperature fusing the ink onto the body; a protective clear glaze was then applied and the item was fired again at a higher temperature. The copper plates were time-consuming to produce initially but could be reused a large number of times, indicating that Hopkins must have commissioned a considerable number of vessels from Turner's. It would appear that all these transfer-printed names were produced from the same copper plate. Although the transfer-printing can be viewed as a technological advance as compared with impression, it is inferior in terms of legibility and aesthetic appearance in this instance. By contrast, the Turner vessels are of noticeably better quality fabric and finish than the other plates in the assemblage made from the same general fabric and the maker's marks can be viewed as a form of quality related branding, similar to Samuel Wilkinson's pipes. This may also explain why Turner's used their full name for the mark whilst other manufacturers just used initials, as by the time these plates were produced Turner's had gained a significant reputation (Hillier 1965) and the name may well have had noticeably positive associations in the minds of consumers. At this time college cooks were semi-independent contractors, responsible for the internal management of the kitchens and the provision of food. They had to supply the ceramics used and it is likely that the cooks and college names on plates were placed there either in an attempt to prevent theft or as a mechanism related to compensation for breakages. As such they can be viewed as part of an 'institutional' archaeology (Evans 1990;Evans and Pollard 1999); the writing must have operated at both personal and institutional levels and the meanings and associations conveyed at the different levels may well have varied. Both the impressed maker's marks and the hand-painted or transfer-printed cook's name were positioned so as to be invisible during use but are visual in nature: even though the impressed marks avail themselves to tactile engagement via the raised and depressed surfaces which form the letters, they are positioned so they would not be felt easily during use. The visual : tactile dichotomy is somewhat simplistic with some types of writing operating in both spheres; it is also worth noting that although the plates were used for serving food the writing may also have served functions at other times such as when the plates were being stored, selected for use, washed, etc.
There were a number of other creamware vessels with the names of College and College cooks, including "Trinity Hall" and "TRINETY H…", "CAI…" for Gonville and Caius College and "B F Tunw…", which can be linked to Bates Francis Tunwell, the Emmanuel College cook (1794-1806). There were also impressed makers' marks "IH" (J. Heath of Hanley c.1770-1800) and "CB" (Charles Bourne of Fenton c. . Whilst documentary sources suggest several routes through which various cooks' plates may have ended up at the Cock Inn, the presence of pieces marked with college names is difficult to explain. The wares may have been purloined, but this is difficult to prove. Some of the marked plates probably ceased to be used for their original college function 10 to 20 years prior to their deposition, indicating that individual pieces and the assemblage as a whole have a 'biography' that must be taken into account in our interpretations of the function and meaning of writing (see below). Very different is a creamware cup decorated with the hand-painted text that probably read "For my dear", as this would have been visible during normal use. Also visible during normal use was an ale measure mark comprised of a crown over the initials "WR" on a stoneware tankard-shaped ale measure jug. The jug was marked in compliance with the act of 1700 for ascertaining the measures used for retailing ale and beer which covered vessels of up to a quart capacity used in inns and other commercial establishments and was in force until 1876 (Binson 1970). This mark can be viewed as part of an "archaeology of regulation" (Egan 2009: 281), as it was mandated by a higher authority. This writing was both a material expression of institutionalised structures concerning regulation and associated practices, as well as actively constituting the physical execution of those practices (e.g. decanting certain types of liquids).
A pearlware jug also found in the Cock Inn cellar shows "The Sailor's Farewell and Return" motif, a common design c.1790-1800 (Lewis and Lewis 2006: 2, 15, 156). Consisting of two scenes, the first depicts the departing sailor and his lass waving goodbye with his ship in the background. The second shows the returning sailor consoling his girl who has wed another in his absence. While there is no writing on this jug, numerous other contemporary vessels decorated with this design do bear writing. The design relates to a traditional folk song and the jug would have brought to mind the words of the song to those knowledgeable viewers who saw it. Essentially when texts become well known they can be evoked on material objects with the text itself being immaterial. This is paralleled by 'literary ceramics' deposited in the 1840s at High Wycombe, Buckinghamshire. These were decorated with scenes from a number Walter Scott's novels, Miguel de Cervantes' Don Quixote and James Thompson's The Seasons (Lucas 2003;Lucas and Regan 2003). These 'literary ceramics' lack text altogether and indeed the selection of images, "suggests, ironically…that the production and consumption of literary images on transfer-printed earthenwares was only successful in so far as such images were relatively independent of their literary reference, or that the literary reference was at least almost universally known" (Lucas 2003: 140).
Within the wider cultural context, this jug and similar 'literary ceramics' with imagery calling to mind certain phrases or verses, can be seen as a kind of material reification of writing, despite its physical absence. Likewise, since not all viewers, whether children or adults, would have been literate, even where writing is present imagery could have also served mnemonic purposes. Depending on the knowledge of the viewer then, writing and image could have served as two different means to the same end, or could have been seen as complementary or overlapping in their purpose. These examples highlight the complexity of the relationships between writing, literacy and oral traditions that should be borne in mind when considering both the physical expression of written meanings and their invisible counterparts.

Sarah Dobson's Planting Pit, F.3010
The third group is from a planting bed dug c.1822-1840 in a garden used by a school run by Sarah Dobson (Figure 5). The dating and composition of the assemblage makes it clear that it relates to the school, whose premises were occupied by Sarah Dobson along with her nieces, a number of pupils who lived at the premises and two servants. Here, writing is only found on pottery, the majority of which consists of manufacturer and pattern names transfer-printed on the underside of the vessels. As the writing would have been invisible when the plates and cups were being used for dining and drinking, this suggests it was not intended to be read frequently. Engagement would have been much more restricted, perhaps only being read during washing up or occasionally when an individual wished to purchase more items of the same pattern. The writing from this assemblage that would have been visible during use includes two children's cups with pink transfer-printed 'moralising' decoration (Cessford 2009: 313-317;Crook et al. 2005: 148;Jeffries et al. 2009: 336-340;Karskens 2001: 74-76). This also included the text "For I have food while others starve or beg from door to door" (Figure 6), part of the song "Whene' er I take my walks abroad". Comprising the second half of one verse from a six-verse song, the text on the cup represents less than 10% of the original poem which appears in the collection "Divine and Moral Songs for Children" by Isaac Watts (1674-1748), a leading early 18 th -century non-conformist hymn-writer, theologian and logician (Argent 1999).
Despite his nonconformist beliefs Watts's work with its straightforward and relatively gentle Christian ideas plus its lilting metre became extremely popular and was frequently reproduced on children's ceramics of the period (Riley 1991: 228-232). The relationship between writing and the  material context of its expression is illustrated particularly well among these finds, especially the impact of the materiality on the content of the writing. The physical form of the cups, with their rounded smaller surfaces, means that they generally have just one or two lines of verse in contrast to the plates with their large flat surfaces which usually bore a whole verse, if not two. There are also differences in the ratio of text and image between plates and cups, the plates tending to be dominated by text, while the cups were often dominated by associated imagery. However, in a planting hole near to the planting bed a fragment of another child's cup was found with part of the text of another work by Isaac Watts entitled Innocent Play. In this case the text seemed to be the dominant decorative feature as there was no evidence of any accompanying image. The text probably therefore included the entire first verse of this three-verse poem.
As well as differences in the ratio of decoration types, an interesting difference between artefact types emerges when we consider embodied practice. The image and / or text on a cup is arranged so that it wraps around the exterior. It is therefore never visible to the user in its entirety and requires rotation for full viewing. Likewise, for non-user viewers the decoration would only be revealed episodically as the cup was filled, drunk from, and otherwise manipulated during the course of use. In contrast, the text on the plate would be wholly visible when the plate was empty, whether on display in a cupboard or as part of a freshly laid table. Yet similar to the cup, its decoration, too, would be partially concealed when filled with food and a process of revelation would ensue as the plate was emptied of its contents. Thinking about these finds and their textual and pictorial decoration in terms of daily practice reveals a complex network of meanings that extends beyond, and therefore require consideration alongside, purely semantic functions.
Returning to the planting bed, found along with the cups and plates just discussed were seven vessels with "Sicilian" pattern decoration (Figure 7), including four plates, two large serving dishes and a cup. Together these can be understood as forming a 'service' , the presence of which can be linked to developing 19 th -century ideas and practices of domesticity and gentility (Fitts 1999). One of these actually bore on its underside the label "Sicilian", which appears to be a pattern name inspired by the gothic novel A Sicilian Romance by Ann Radcliffe (1764-1823) published in 1790 (Coysh and Heywood 1982: 338;1986: 183). In this respect these vessels, too, may be understood as a form of 'literary ceramic' , connected to changing perceptions of fiction and its accompanying illustrations. Their use may be understood as mediating ideals of the picturesque and suitable subjects for transfer-print patterns (Lucas 1993). In the case of the Sicilian pattern the image of the Mediterranean scene could have functioned independently of the novel, simply as a picturesque view. However, Radcliffe's books assert traditional moral values such as honour and integrity while making strong political statements concerning the oppression of women in patriarchal society. Given the composition and nature of the 'household' living at the premises, it is likely that Sarah Dobson was responsible for purchasing most, if not all, the ceramics and the ideas expressed by Radcliffe may well have appealed to her.
The cups associated with Isaac Watts bear only a small portion of the original songs that they derive from, whilst the Sicilian pattern vessels have no text from Radcliffe's book. This again reinforces the point that the translation between paper and other mediums was successful when the text and / or images were relatively independent of their literary reference or the literary reference was almost universally known.
The writing associated with Sarah Dobson's school is particularly significant as it relates to an institution where the process of writing itself was central to the establishment, as demonstrated by the fact that 17 of the 18 slate pencils recovered during the excavations can be associated with this school. The pencils also suggest that many members of the household who used the vessels discussed would have been able to read the writing present on them, something that is crucial to consider when the materiality of writing is being discussed. Although the cups with texts were used by children, it was Sarah Dobson who controlled their selection and it is in the light of her world view that they must primarily be interpreted, although it should be recognised that other members of the household may have read very different meanings into the texts. Any such interpretation must hinge upon whether the texts present in the assemblage functioned independently, or whether they derived their significance from the larger works that they referenced.

Thomas Wicks's Soakaway, F.6412
The fourth group of writing-bearing objects comes from the backfilling of a brick-lined soakaway dating to the early 19 th century which yielded finds relating to the occupancy of Thomas Wicks, a cook at nearby Emmanuel College (1807-1852; Figure 8). One of the most common forms of writing in this context was the impressed maker's marks on creamware pottery, including  Wedgwood (MNI 12) and Turner (MNI 9). As with the earlier Turner marks, these also appear to be a form of quality-related branding. A number of 'services' of related vessels were identifiable, one of these included a sauceboat with "T.Wi…" hand-painted in gilt indicating that this must have belonged to Thomas Wicks. While some of the vessels in this service are incomplete, others are whole enough to ascertain that they were not marked with Wicks name. This raises the question of why some vessels in a service were marked and others were not. One possibility is that the sauceboat was a relatively expensive item, which made it worth marking. Alternatively it may have been linked to the rather different role of the sauceboat, which would have been used by all the diners at the table, rather than just one individual as a plate would. Among the soakaway finds there was also a wine bottle with a glass seal marked "E G 1770" relating to the individual for whom the bottle was manufactured and the year in which this occurred. No one with these initials can be linked to the property and as Thomas Wicks was only baptised in 1774, four years after the bottle was manufactured, it is unlikely that the writing had any special meaning for him personally. This seal is better executed than the Pyrmont bottle discussed previously, although it is slightly misaligned. The initials "E G" are clearly legible, unlike those on the Pyrmont Water bottle; indeed if such a personal seal had been so poorly executed during manufacture as the Pyrmont Water one, it is likely that the bottle would have been rejected as the seal would have been rendered pointless.
Additional finds included nine used clay tobacco pipes bearing the mark of Cambridge-based pipe makers James and / or Ann Pawson. James inherited the pipe-making premises and business of his uncle-in-law Samuel Wilkinson, who has already been mentioned, in 1786. In 1813 James's wife, Ann, succeeded him, remaining active until 1823. Initially James Pawson marked the pipes in a similar manner to Wilkinson, with ornate curvilinear decoration and the writing "J•PAW-/ SON, Cam-, Bridge" or "JAS.PAW. / SON, Cam•/ Bridge", which was impressed using a roller stamp before the pipe was fired. At some point either James or Ann Pawson switched to the stem mark "PAWSON CAMB" enclosed in a circle. All these marks were similar in terms of location and size to those on the pipes produced by Wilkinson that have already been discussed. They were relatively small and would have been obscured by the smoker's hand whilst they were being used, so engagement during use would have been primarily tactile rather than visual and it would only have been possible to read the text by deliberately examining the pipe. By continuing with the same style of stem marks as his predecessor, James Pawson was perpetuating a branding tradition and with it the business 'goodwill' or reputation built up by Wilkinson. The transition to the different style of mark was probably prompted by the roller stamps that the Pawsons used becoming so worn that they were un-useable, by which time that style had gone out of use and such roller stamps were no longer being produced.
The instances of writing from Thomas Wicks's soakaway are of themselves generally unremarkable, indeed they were similar to examples deposited decades earlier. In some respects this is key to their functioning as these are essentially repetitive forms of writing where similar texts in similar forms had been occurring in similar locations on similar types of artefacts for periods that often exceed the lifetime of a typical individual who read them. Although the texts themselves are often relatively novel, giving the name of a particular maker or owner, the act of reading is one embedded in daily social practice and memory (cf. Hodder and Cessford 2004).

Barrett's Ceramic Retailers, F.4060 and F.4106
The next assemblages are rather different as they relate, in part at least, not to items owned by a particular household but to the stock of a business that sold a range of material including items with writing on them. Between 1882 and 1885 the Barrett family, retailers in china, earthenware and glass, reorganised the rear area of their premises at No. 25 St. Andrew's Street. Two separate features were backfilled during this period, a rectangular sunken structure (F.4060; Figure 9) and a cellar (F.4106; Figure 10). The finds in these features appear to represent a mixture of contemporary finds related to the Barrett family business, plus some older items linked to the clearance of the garden. These latter artefacts include the remarkable find of a large and nearly complete Martaban storage jar. The vessel, originally from Southeast Asia, was stamped with the Chinese symbol for the Boar (inoshishi), one of the 12 years of the 'Sheng xiao' commonly known as the Chinese Zodiac. The symbol is relatively small in relation to the overall size of the vessels and quite discrete due to the 'textured' nature of jar. There does not ever appear to have been an active trade in these jars with Europe, instead they seem to have been used occasionally as containers for water, oil and other substances on board vessels and thus made their way to Europe. Occasional pieces of Martaban have been found in 17 th -19 th century contexts in Britain before, but usually only as single or small numbers of sherds. The much more complete example from Cambridge probably arrived in Britain in the same manner as the others, but it may have acquired a kind of curio status subsequently, perhaps by a member of Emmanuel College given its find location on property occupied (c.1833-1847) by the college butler Charles Burbage. He seems to have put the jar to use in the garden, perhaps as a water container judging by the pattern of limescale on the vessel. The boar symbol relates to the year that the jar was manufactured. Its small size relative to the large object and given the probability that it rapidly became unintelligible once it moved away from its area of production make it likely that many of the individuals who came into contact with the jar were probably unaware of its presence, could not have understood the symbol, and may not even have recognised it as a form of writing.
After the alphabetic script used to write the English language, Chinese characters are the most common form of writing found at this site. They occur on both Chinese imports and local imitations. The majority of instances appear on local British ceramics and include blue and white transfer-printed designs of a Chinese style, yet the 'texts' , and indeed even individual characters, are frequently gibberish. This is part of a much wider phenomenon, whereby from the 17 th century onwards Chinese artistic influences had a huge influence on British culture leading to the development of the 'Chinoiserie' style (Honour 1961;Impey 1977).
An approach that is increasingly being applied to 18 th -20 th century archaeological material (e.g. Dellino-Musgrave 2005;Mytum 2003) is the concept of artefact 'biographies' (Lucas 2005). This is particularly apposite for the Martaban jar boar symbol, a symbol that would have been intelligible to many in its production context, but probably not to most sailors, merchants or others on board the European vessel that transported the jar to Britain, or those otherwise involved in its transport to Cambridge. Unlike in many parts of the world, there is no evidence for a Chinese community in Cambridge and only limited evidence that anyone could have read the symbol. Cambridge University Library obtained its first Chinese book in 1632, but there was no official academic interest in Chinese until Sir Thomas Francis Wade (1818-1895) was appointed the first holder of the Chair of Chinese (1888-1895), some three years after the Martaban jar was deposited. A nearby department store, Robert Sayle, had strong links with Shanghai and Hong Kong possibly as early as c.1860 and certainly by c.1870-1872, continuing until the early 20 th century (Sieveking 2004: 32-33). It is likely that some members of the Robert Sayle business, including members of the Sayle family itself, would have been able to read Chinese characters.
Based on present evidence, the ability of most viewers / users of the jar to read the writing was limited. Nevertheless, many would have been able to recognise the type of script which may have served as a reminder of its biography and led to it being kept, gifted and reused for a time, especially when it remained in a collegiate context. Its role in Burbage's garden is less clear, although it is possible that this was a place of social display where the jar signalled Burbage's role as a relatively powerful college servant with access to unusual objects from rare lands and, perhaps more significant from the point of view of those viewing it, the ability to appropriate material that was usually restricted to the local social elite. Several glass bottles including two with seals marked "EMANUEL / COLLEGE" and "EMANUAL / COLLEGE" dated c.1820-1860 were also probably brought to the site by Burbage, the college butler, and re-used as building material as several have traces of mortar on them. There are seven known wordings of Emmanuel College seals ("EMANUAL COLLEGE", "EMANUEL COLLEGE", "EMMAN. COLL", "EMAN. COLL", "COLL. EMAN", "EMM. COLL" and "EMANUEL COLL" [Morgan 1977: 70]), and it is likely that each distinct seal represents a separate order from a glassmaker (Banks 1997;. The Emmanuel College seals are all relatively well executed and carefully aligned, contrasting with the earlier glass seals discussed. There are no technological reasons for this improvement and the most likely reason is simply that the college was a demanding client and the high social status of its members meant that they required, and were in a position to enforce, high standards. Archaeological finds linked to the Barrett family business date largely to the 1870s. There are two ceramic water jugs, decorated in black and green respectively, with geometric bands and classical hunting scenes and transfer-printed marks on the base consisting of a diamond-shaped registration mark containing a mixture of letters and numbers, the pattern name "TRAJAN", and the maker's mark "P&B". Between 1842 and 1883 some pottery was marked with a diamond shape printed or impressed on the base, a symbol relating to the British Patent Office Registry of Designs. In the case of the Trajan jugs, this indicates the class or type of material (IV, clay ware), the bundle or number of items included in the registration (5), and the day (28 th ), month (H, April) and year (F, 1873) of the registration. Important for the 'archaeology of regulation' , the information conveyed in the diamond registration mark was so highly codified that it was likely that only a small percentage of the vessel users would have understood it. Moreover, it would have been largely invisible during normal use. Similarly the maker's mark "P&B", Powell and Bishop, who were in partnership in Hanley Stoke on Trent (1876-1878), was probably equally cryptic. The pattern name "TRAJAN", the impression of which was incompletely executed, may evoke the Roman emperor Marcus Ulpius Trajanus (52-117 ad), perhaps via a literary allusion inspired by Pliny the Younger's Panegyric Panegyricus Trajani of 100 ad where Trajan is praised for his interest in hunting (81.1-3;Bennett 1997: 66).
At some level the decoration on the Trajan pattern jug was designed to appeal to individuals who considered themselves cognisant and appreciative of the aesthetics of classical art, yet by this date such influences were no longer restricted to the social elite and the jug is by no means an exclusive product. The name Trajan encodes meaning based upon knowledge of a relatively exclusive text, which members of the educated elite would be aware of but which probably escaped members of the lower classes who emulated them.
The Barrett's brick-lined cellar (F.4106) included at least four identical white stoneware teapots, each bearing the maker's name "COPELAND" and diamond registration mark impressed into the base manually using a stamp before firing. This diamond mark is similar to the codified information on the Trajan jugs just discussed, indicating that this was a clay ware (class IV), the bundle or number of items included in the registration (2), and the day (24 th ), month (H, April) and year (T, 1867). The layout here is slightly different from the later Trajan jugs as the organisation of diamond registration marks changed in 1868. In contrast to the transfer-printed Trajan pattern, the diamond mark and maker's name are embossed and therefore three-dimensional. However, rather than this reflecting a specific choice related to the writing, the technique used for the registration mark in both cases relates to the technique used for the overall decoration of the item. Embossing and transfer-printing both have their strengths as techniques when used for writing, transferprinting with its greater colour contrast is generally easier to read whilst embossing is more durable. Material expressions of writing need consideration in relation to the context of manufacture since their appearance is linked to wider technological practices.
The same cellar assemblage included four highly fragmentary children's mugs also bearing writing. A purple transfer-printed alphabet cup has the text "B IS FOR BUFFALO, C IS FOR CAT" running around the bottom of the side of the cup, although the vagaries of production mean that "BUFFALO" is partly missing. Another purple transfer-printed cup fragment has the text "LOVE YOUR ENEMIE…OVER…", the first part of which derives from Matthew 5:44 "But I tell you: Love your enemies and pray for those who persecute you". Rather appropriately given its Biblical origin, this text is made to appear as if it is in a book. A third purple transfer-printed scene of children playing has the text "...ATHER THOU A…", the whole cup would have shown "MY FATHER THOU ART THE GUIDE OF MY YOUTH" (Jeremiah 3:4) and "THY WORK IS A LAMP UNTO MY FEET" (Psalm 119:105) (Riley 1991: 248-249). A black transfer-printed example has the text "REME[MBER] / THE SABBATH / DAY / TO KEEP IT / HOLY" from the fourth commandment (Exodus 20:8). This text is shown carved on a stone object, alluding to the 10 commandments Moses received on two stone tablets from his god. Children's ceramics promoting piety and virtue were common in Britain in the 19 th century (Riley 1991: 226-259) and these types were frequently given as Sunday School prizes for good attendance and achievement (Riley 1991: 248). The selection and extraction of texts on these cups from a longer paper-based piece of writing mirrors the practice already discussed of only including a small proportion of an original text, and implies that either the audience was familiar enough with the original text to understand the meaning of the fragment used or that its meaning was clear enough to function independently.
What is particularly interesting here is that, although the cups from this cellar on which the writing occurs are virtually identical in size, form and material, the way the writing is presented in terms of its location, font used and integration with images varies considerably, effectively relating it to its Biblical origins. This writing is also strongly linked to children, a phenomenon present in several other features discussed here. Items of material culture linked specifically to children are relatively rare in terms of the overall assemblages from the site, but writing occurs on a much higher proportion of these than on items associated with adults. Also significant is that the location, scale, and other features of its material expression indicate that writing that is intended to be visible, read, and its meaning well understood. The wider social context of this is that the concept of childhood changed markedly during the 19 th century and the middle class in particular came to view it as much more separate and distinct from the adult world. This involved numerous changes, particularly with regard to education -something which is relevant to explaining in part the significance of the materiality of this particular writing, especially its clear visibility.

The Robert Sayle Cellar, F.4027 and F.4127
The final assemblages I examine in this chapter date to 1913-1925 when a double-roomed cellar (F.4027 and F.4127) used by the department store Robert Sayle was backfilled (Cessford 2012; Figure 11).This is the latest feature discussed and by this time writing had become even more common, appearing on a wider range of materials and object types. The majority of the items with writing on them appear to be the possessions of members of staff who lived at a dormitory on the premises, so in some sense they are a 'household' group albeit one much larger and very different from most of the others that have been considered here. There is relatively little that relates directly to the business itself, although some plastic oval-shaped furniture fittings, probably for drawers, were embossed with "R.SAYLE & CO. / CAMBRIDGE" representing a form of corporate labelling.
A range of ceramics, whole and fragmentary, was also found here, many of which bore writing of one type or another. The majority of these were manufacturer and pattern names rendered using the transfer-print method and many of which were poorly executed and are illegible or only semi-legible. Registration marks continue to be used, but from 1884 onwards the diamond shaped marks ceased and were replaced with a consecutive numbering system, such as "Rd.No.510607" which designated The Pompadour pattern. Whilst the earlier diamond system had been heavily codified it potentially conveyed some information to those with an understanding of the system such as pottery retailers, whereas now the consecutive numbering system was an abstract number that in isolation meant virtually nothing. Many of the ceramics from this feature are also marked "ENGLAND"; the McKinley Tariff Act of 1890 meant that all imports to the USA needed to carry the name of the country of manufacture and in practice this meant that much material for domestic use was also marked.
Another item from the department store cellar was an alphabet plate with raised moulded letters on the rim running from A to Z around the rim. Within this was a blue transfer-printed central scene of a group of dolls surrounded by the symbols of the manual sign language alphabet. On the back is a transfer-printed mark "R D .N o .426673 / H. AYNSLEY & C o / LONGTON / ENGLAND". The moulded letters on the rim are essentially an 'off the peg' element, used for a wide variety of children's plates. It is apparent in this case, as in most instances, that when the transfer-printed design was added during a later stage of the manufacturing process, no attempt was made to orient the two designs so the layout of the letters corresponded. Sign language is a visual-gestural language and British Sign Language as it exists today probably originated in the 18 th century. From the 1860s sign language fell out of official favour and oralism, which uses spoken language consisting Sayle & Co label and sign language plate. Photographs by Dave Webb and author, drawings by Vicki Herring.
of lip reading, speech, the process of watching mouth movements and mastering breathing technique, was strongly promoted. This became official policy after 1889 and the late 1890s and the early 1900s were the heyday of the oralist approach and sign language was heavily discouraged. It is unclear if items such as this plate, whose design was registered in 1904, were produced for deaf children specifically, or if they were part of a campaign and were used by other children (Riley 1991: 120-121). In any case given the dominance of oralism at the time, such plates can in some respects be viewed as artefacts of 'resistance' (Frazer 1999), with members of the deaf community resisting the perceived wisdom of those exercising control over social and education policy. In some sense resistance is the opposite of regulation, although as this plate bears both sign language, which was disapproved of by the authorities, and also a Government imposed registration mark, the writing is comprised of elements from both on a single item. Evidence of 'resistance' is much rarer than that of 'regulation' , although this may be due in part to the former often being more subtly expressed. One example of this is at a property where the name of the owner, Emmanuel College, is prominently displayed on the frontage whilst the tenants' initials were placed in a more discrete location in a manner that suggests illicit behaviour (Figure 12). By the early 20 th century the presence of writing on artefacts had become much more common -a trend that is apparent throughout the 19 th century. With increasing frequency it appears that in this text-saturated world most instances of writing were expressed in a way that suggests they were not meant to be read. Increasingly writing was placed and designed to be as unobtrusive as possible when the item was being used for its primary purpose. Such texts were either mandated by the authorities or were intended to be read infrequently and for often rather abstruse purposes. Running counter to this in a few instances, such as the sign language plate, the text maintains its visibility and remains central to the use and social meaning of the item on which it appears. In such cases the writing often becomes increasingly prominent, to counteract the effects of its textsaturated world.

Conclusion
The aim of this study has been to demonstrate the importance of accounting for writing as material and as part of individual and social practice. The 'feature group' approach has an important contribution to make to the study of the materiality of writing of 18 th -20 th century Britain. Whilst each feature contains its own narrative, the aim of analysis at the scale of the 'feature group' was not solely to consider the individual assemblage in isolation but as a starting point for discovering larger patterns. The 'elephant in the room' is the fact that what survives archaeologically is only a subset of past writing materialities. The most common medium for writing in 18 th -20 th century Britain, probably by several orders of magnitude, was undoubtedly paper. This did not survive in any of the feature groups discussed, but in those rare archaeological instances where paper does survive from this period it vastly outnumbers writing on other materials (e.g. Crook and Murray 2006). Archaeologically, this scenario where the dominant medium for writing is the least likely to survive is paralleled in other cultures (e.g. Waal 2011). It is clear that in 18 th -20 th century Britain writing on paper was regarded as the norm, with all other materials viewed as secondary. Indeed many of the examples discussed are derived either directly or indirectly from writing on paper. Examples of this include the cup with text from a published poem (Figure 6), the Sicilian pattern vessels which rely upon a novel for much of their meaning (Figure 7), the Trajan pattern jug whose imagery relates to the Panegyricus Trajani (Figure 9) and children's mugs which derive from and in one case actually depict the bible (Figure 10), the book par excellence of Britain during this period. In other cases the writing makes no sense without the existence of writing on paper; for example registration marks are meaningless without both the enabling Act of Parliament and the supporting 'paper trail' of the individual registration process (Figures 9-11). The archaeological preponderance of writing on what might be viewed as secondary expression raises significant questions about what studying its materiality can tell us. The use of writing, both on paper and other materials, becomes increasing prevalent over the period in question. As a counterpoint to this phenomenon the texts and their meanings on materials other than paper in many cases becomes less visible -apparently not meant to be read by those who are otherwise consumers of the objects. Furthermore, much of the writing relates to regulation and there are also repeated links to the education of children and both commercial and institutional branding. All of these phenomena, as well as the crucial underpinning factor of increasing literacy rates at this time, must therefore be situated in the context of major historical processes of the period such as modernity, capitalism and consumerism. The archaeological study of writing as material practice at the scale of feature groups sheds light on how particular households in specific temporal, spatial and social milieux interacted with those forms of writing that survive. At a broader level these specific examples attest to the development and spread of a text-saturated culture, a phenomenon which is inextricably intertwined with the major historical processes affecting 18 th -20 th century Britain.