Vivian Cook, Newcastle University
Punctuation Web   Language of the Street

In M. Pawlak and L. Aronin (eds.), Essential Topics in Applied Linguistics and Multilingualism, Springer International Publishing Switzerland (2013), 267-290.[Note the analysis is changing a bit in the book in progress and further punctuation stats etc are found on via Punctuation website]

The city streets are alive with written language. It’s impossible to walk for a few yards without seeing the numbers of houses, the names of restaurants, for sale signs on buildings, no parking notices, signs for hydrants, posters, and any of a thousand street signs we take for granted every day of city life. The street is a complex public area in which people move, live and have relationships, at the core of city life (Jacobs, 1961; Hall, 2012). This paper discusses the punctuation component of the language of the street, based on the approach developed in Cook (2013), which draws on linguistics, writing system research and linguistic landscape research to analyse the street signs in one urban milieu in relationship to those who create them and those who read> them. The paper is an exercise in social semiotics (Scollon & Scollon, 2003) focussed on visual signs with linguistic meaning.

Punctuation concerns the visual aspects of a writing system that convey information by symbols other than the letters of the alphabet in sound-based writing systems like English or the characters in meaning-based systems like Chinese. The paucity of research into punctuation means spending some time establishing the facts of English punctuation in general to bring out the unique features of the punctuation of the street. English punctuation has a rich history of prescriptive discussion and analysis, reflected in the perennial popular debates about the ‘correctness’ of some use of punctuation, such as the so-called greengrocer’s apostrophe seen in <Panini’s>, found on every street in the land. Yet little modern empirical research has been devoted to punctuation, compared say to the many studies of texting. Indeed most linguistic descriptions of English have ignored punctuation: it is for example mentioned only twice in the 1204 pages of the comprehensive Longman Grammar (Biber et al, 1999) and is excluded from the frequency accounts for the British National Corpus (Leech, 2013). Punctuation is evidently not awarded the status in the description of written grammar that intonation is given in the spoken grammar.

However, virtually all linguistics makes surreptitious use of punctuation. Take any example sentence from a paper on syntax such as:

John said he was looking for a cat, and Bill did too. (Chomsky, 1995, 126)

This is separated into words by eleven word spaces, marked for class of noun by capitals <John> and <Bill>, divided into two clauses by a comma, and delimited as a sentence by its final full stop, as one can see in its unpunctuated form:


Most of the data presented as examples in linguistic papers have been pre-analyzed through punctuation for certain structures and word classes, even those presented in phonetic script– with word divisions.
Cook (2013) looked at all the meaningful signs from two streets in Newcastle upon Tyne, Stowell Street and Leazes Park Road, between March and June 2012, ranging from restaurant fascia boards to manhole covers. The aim was total coverage of one area at one time, in the manner of the classic study of Las Vegas by Venturi, Brown and Izenour (1972/1977). More details of these two streets and of the demography of Newcastle upon Tyne are given in Cook (2013); they are mixed usage inner-city streets of shops, restaurants, pubs and clubs with primarily nineteenth century buildings. The present paper develops one area touched on in that paper, namely punctuation in these street signs in English, i.e. excluding the bilingual signs. This reduces the total to 296 signs. The omission of bilingual signs is partly because punctuation is to some extent universal in Western writing systems, apart from differences in the form of some punctuation marks (Nunberg, 1990, 10), but also because there appears to be little description of it in the Chinese writing system, the main other system represented in our signs.

The conventions followed here are that examples of written language are enclosed in arrow brackets <Big>, examples of spoken language in phonetic script are given in slants /big/, while linguistic example are given in italics hullo. Since the data is visual, many examples have to be reproduced here, the only editing being altering the appearance to enhance legibility on the page. In the examples in the text the typefaces of the letters are approximated but the originals
convey them best.


Traditionally the nature of punctuation has been interpreted in two parallel ways as rhetorical versus grammatical punctuation (Parkes, 1992, 4), called Cook (2004). correspondence punctuation and grammatical punctuation.

- correspondence punctuation. The overall idea is that punctuation marks correspond to something in the stream of spoken sound. Punctuation marks such as commas < , > and full stops (periods) < . > are claimed to correspond to different lengths of pause in reading aloud.

One of the earliest systems was that of Puttenham (1589):

… the auncient reformers of language, inuented, three maner of pauses …The shortest pause or intermission they called comma …. The second they called colon… [which] occupied twice as much time as the comma. The third they called periodus, for a complement or full pause…

This is virtually identical to the advice to be found online today:

Where you think a reader should make a major pause (draw breath), use a full stop. Where you think a reader should make a smaller pause, use a comma... (Wilson, 2007)

Punctuation marks are also taken to correspond loosely to a few intonation patterns for reading aloud. Typically a question mark < ? > implies low-rise tones for yes/no questions <Are you all right?> and an exclamation mark < ! > implies rise/fall tones <Well!>, while commas may be used for low-rise tones in lists <A, B, C, D>.

- grammatical punctuation. Punctuation marks also show the grammatical structure of the sentence. An early account is found in Lowth (1762), an influential though much derided source. While primarily an advocate of correspondence punctuation, he nevertheless stated:

The period is the whole Sentence, complete in itself, wanting nothing to make a full and perfect sense and not connected in construction with a subsequent Sentence. The Colon, or Member, is a chief constructive part, or greater division, of a Sentence … Commas, or Segments … are the least constructive parts of a Sentence or Member, in this way of considering it; for the next resolution of it would be into Phrases or Words.

This can be readily transformed into the links between the syntactic rank hierarchy and punctuation presented in Halliday and Mattheisen (2005):

- sentence, shown by an initial capital letter and a full stop;

- subsentence, marked by a colon, semicolon or comma;

- word, bound by spaces;

- letter.

One difference is that clause and phrase are collapsed into a single category of subsentence, rather than using the rank of clause employed in spoken grammar.

We also need to take on board a distinction between lexical sentences and text-sentences (Nunberg, 1990). The lexical sentence used in the spoken language is often described as structurally complete, meaning that it conforms to Bloomfield’s (or indeed Lowth’s) definition of a sentence as a grammatically independent form (Bloomfield, 1932, 170). Biber et al (1999, 202) prefer the term ‘independent clause’, defined as ‘not part of any larger structure but it may contain embedded clauses or be coordinated with clauses on the same level’. A text-sentence, however, is a unit ‘of written text customarily presented as bracketted by a capital letter and a period’ (Nunberg, 1990), that is to say, it is defined by its punctuation. <The dog bit the man.> is a prototypical example of a lexical sentence, while <The dog.>, <Man.> and <Bit.> are acceptable text-sentences. A text-sentence is then anything bounded by a capital letter and a final punctuation mark, namely < . ! ? >, thus encompassing on the one hand the gargantuan sentences of J.R. Tolkien with many clauses linked with commas, colons and semi-colons, and on the other verbless phrases or single words, found for example in Angela Carter’s novels (Cook, 2004).

According to Halliday (1985), the link between pausing and punctuation marks the common coincidence of phrase structure and tone group boundaries. Support comes from Steinhauer (2003) who showed through ERP measures ‘Both prosodic boundaries and commas elicit the same brain response reflected by the CPS [Closure Positive Shift]’.

The rival approaches of correspondence and grammatical punctuation have always had their proponents. The language of the street, however, conveys its meaning entirely through visual means and is not intended to be read aloud; it is quintessentially written language without a spoken equivalent. Traffic signs such as yellow lines on the road surface cannot be read aloud, though they can be paraphrased in speech and in written regulations. Hence the correspondence interpretation of punctuation is not relevant to the language of the street. The rest of this paper deals with punctuation for providing syntactic information in the form of text-sentences and written grammatical units, not spoken ones.


To provide a touchstone for the distinctive punctuation of the street, it is necessary to briefly describe the central features of English punctuation, called by Nunberg (1990) ‘genre independent’ punctuation. These are described in such prescriptive guides as Trask (1997), Carey (1960), Todd (1995) and McCaskill (1998), who provide little empirical support for their pronouncements other than example sentences. This will be called here ‘standard’ punctuation; it is concerned with language as an abstract entity ruled by authority rather than language as a set of sentences that have been uttered (Cook, 2010). The overall issue is how punctuation indicates grammatical units at different levels of the phrase structure of the sentence. Frequency information on punctuation marks will be discussed below.

Table 1. English grammar and standard punctuation

Table 1 displays a thumbnail sketch of the main links between grammatical units and standard punctuation. This amplifies Halliday’s set of punctuation marks with double and single quotation marks < “ ” ‘ ’ > and adds the use of capital letters and parentheses (brackets) (< ( ) >. Boundary marks occur either at the end of the grammatical unit, for example exclamation marks, or in pairs before and after it, commas and dashes. All punctuation marks are followed by a space before any succeeding text, except for < ’ > and hyphen < - >; usage with dashes varies according to the type of dash (en dash < – > and em dash < — >) and their particular uses. Most examples here are quotations from Slavoj Zizek, often from interviews; the remainder are from the sign corpus.

- A sentence is delimited by the boundary marks of an initial capital letter and a final full stop, question mark or exclamation mark < . ? ! >.

<I’m an old fashioned continental European!>

As Nunberg (1990) points out, this delimits a textual sentence, not the prototypical lexical sentence with a compulsory verb etc, as seen in the verbless sentences in:

< Humanity? Yes, it’s OK – some great talks, some great arts. Concrete people? No, 99% are boring idiots.>

- A finite clause may be separated from its neighbour by a boundary colon, semicolon, dash or comma < : ; – , >:

<If you have a good theory, forget about the reality.>

- A non-finite phrase can be separated from its neighbour by a boundary dash or comma:

<Lately we have been doing quite a bit – intervening in foreign countries and destroying the environment.>

Sometimes the phrase has a pair of marks on both sides, whether commas or dashes:

<You could say, in a vulgar Freudian way, that I am the unhappy child who escapes into books.>

The comma, sometimes a stroke < / >, is used to separate structurally equivalent items in lists:

<I am rather perceived as some dark, ominous, plotting, political manipulator, a role I enjoy immensely and like very much.>

This applies in particular to postal addresses:

<Churchill Road, Bicester, Oxfordshire, United Kingdom OX26 4XT>

Another common use is to separate initial adjunct phrases or words from the subject of the sentence:

<Here, I violently disagree.>

- A word is separated from its neighbour by boundary spaces on both sides, here underlined to make them visible:


A word may be joined to its neighbour by a relational hyphen, yielding a compound word:

<The re-focus on the perpetrator’s traumatic experience enables us to obliterate the entire ethico-political background of the conflict.>

Words that are proper nouns or adjectives and a small set of other nouns like Monday start with an initial capital letter:

<… apart from left-radical Keynesians like Paul Krugman, with whom I’m sympathetic…>

- Some morpheme suffixes are joined to their neighbours by a relational apostrophe:

<Today’s racism is precisely this racism of cultural difference.>

- Some omitted letters are replaced with an apostrophe:

<I’m an old Hegelian.>

Other abbreviations have a final full stop:


There are semi-grammatical uses of punctuation that do not relate to the ranks of the sentence, such as quotation marks. In British style, the single quotation mark < ‘ ’ >, in American style, the double quotation mark < “ ” >, signify distancing the speaker through oratio recta, i.e. direct quotation of another person’s words, and can enclose any grammatical unit from the sentence down to the word. Quotation within quotation uses the opposite type of quotation mark for the style, i.e. double marks for British style, single for American. ‘British’ and ‘American’ are labels for styles that are used globally rather than only in the UK or the USA (Cook, 2004).

<‘I hate the position of “beautiful soul”, which is: “I remain outside, in a safe place; I don't want to dirty my hands.”’>. [British style quotation marks added].

Similarly a pair of brackets may be used to enclose virtually any grammatical unit.

<What is really hard for us (at least in the West) to accept is that we are reduced to the role of a passive observer…>

As an aside, there are no analogues in spoken language for capital letters, word-spaces, hyphens and apostrophes. These punctuation marks associated with words, morphemes and letters are thus purely grammatical punctuation. Quotation marks are also unique to the written language; as Nunberg (1990) points out, the rare spoken use of Quote … Unquote is very different from the written quotation marks, as is the teachers’ stylised gesture of holding up two fingers of both hands and wiggling them, baffling to many students.

Two uses that do not conform to the conventional phrase structure of the sentence. One is the use of < . > as a separator between numbers <3.147>, money <£3.14>, web addresses <> and times <13.47>. This concerns numerical rather than linguistic structure. It needs to be mentioned because of its high frequency in the language of the street for opening times, notices, prices etc. It will here be called the numerical stop.

The other semi-grammatical use is the dash showing extent from A to B <1939–1945> and <Open 9–5>, equivalent in meaning to to, as spelled out in one sign <Monday to Friday>. Here a punctuation mark has both lexical and grammatical meaning, as described by Nunberg (1990) for colons in sentences like <Man proposes: God disposes.>. It will be called the fromto dash here.


The functions of street signs can be roughly divided into four systems (Cook, 2013): locating, informing, controlling and service. These will be now be described separately before returning to more general points. The four systems interact with the six roles that people have in relation to signs: licensor, owner, author, writer, addressed reader and unaddressed reader (Cook, 2013). This tries to give a broader account of the many relationships of writer and reader than the topdown and bottom-up dimension typically found in linguistic landscapes research (Ben-Rafael et al., 2006).

1a.  1b.

  1c.  1d.

Set 1. Locating

Locating signs identify the street and its buildings indexically, as seen in the signs in Set 1, called by Scollon & Scollon (2003, p.146) ‘situated semiotics ... any aspect of the meaning that is predicated on the placement of the sign in the material world.’ The meanings of locating signs are false if they are moved to a different location. The number <18> (Fig 1a) would not be true if it were lying on the pavement or attached to the house next door. Austin (1962) laid down felicity conditions for performative utterances such as the speaker having the right role, i.e. a judge pronouncing sentence. The felicity conditions for locating signs require not only the right licensor and owner but also the right location. Some signs, like brassplates for businesses, have to be displayed by law in the Companies Regulations (2008). Their readership is passers-by, drivers, postal workers etc – anyone who needs to know the precise identity of a particular spot. Fig 1b <BARKER & STONEHOUSE> is a locating sign for a shop, giving nothing but the name, indexical to the building it appears on, made of metal with raised capital letters. Scollon and Scollon (2003, p.153) see this as exophoric reference in which language refers to a particular visible physical object. Fig 1c <LEAZES PARK RD.> is a typical modern English street sign, with raised capital letters and an abbreviated <RD.> with full stop; the physical object to which it refers is then the entire street, presumably extending to the point where a different street-name sign is placed.

Grammatical punctuation is virtually absent from locating signs and they have no sentence-final full stops. The word-rank possessive -s’ apostrophe is found, for example in Set 1d <ROSIE’S BAR>, but is often absent, as in <KATHERINES FLORISTS> and <GREGGS>. Many signs simply assert the name or number of the premises, as in Fig 1b <BARKER & STONEHOUSE> and Fig 1a <18>. For many, as in Fig 1b, this is proclaiming not just location but also ownership. This can be called a Locator. Most Locators consist of one or more noun phrases with proper nouns, in Fig 1b connected with ampersand <&>, frequent in street signs.
When the Locator is too long to fit on one line, it is divided by line-breaks, as in Fig 1b, which coincide with word divisions.

2a.   2b.
2c.   2d.
Set 2. Locating Signs

A Locator such as Fig 2a <KING NEPTUNE> can be accompanied by additional information – an Expander such as <SEAFOOD & PEKING RESTAURANT>. An Expander is typically another noun phrase or pair of phrases, separated by a line break from the Locator, such as <A Warm Welcome to All>, below the Locator <ROSIE’S BAR> in Fig 1d. The line-breaks within a Locator or Expander mostly correspond to grammatical divisions but not necessarily. For example Fig 2b belies the underlying phrase structure by implying (kingswalk dental) (implant practice) rather than (kingswalk (dental implant) practice). The apostrophe used with plural nouns occurs often in Expanders, as in <Exceptional Florist’s>, <Panini’s> and <under 7’s>. (Perhaps it should be noted that the use of apostrophes with single letters and numbers is advocated by Oxford Dictionaries ) In both Figs 2a and 2b the Locator is in ‘conservative’ serif, the Expander in modern sans serif (Cook, 2013).

Fig 2c also contains an Identifier giving an address, whether web address <>, street address <50 Leazes Park Road> or phone number <Tel: 0191 236 6622>. These have specific address punctuation such as numerical stops, commas and colons <Tel: …>. Finally Fig 1d has a List structure of items for sale <Cask Conditioned Ales …> with no punctuation, to be discussed later. The visual relationship between Locator, Expander and Identifier is not necessarily from top to bottom of the sign. Fig 2c has an Identifier <58 Leazes Park Road> above a Locator <LMR> and Expander <Recruitment Consultants>, followed by more Identifiers such as the phone number. The Locators in Fig 1d, 2b and 2c are more prominent by virtue of being larger or bolder or having a more distinctive font, not just by position. The Locator is often all capitals (Figs 1b, 1c and 2a) or all lowercase (Fig 2d) rather than word initial capital letters.


  3a.   3b.  3c.
Set 3. Informing Signs

Informing signs provide factual information such as opening times, the availability of goods and requests for planning permission. Informing shades into job offers, for sale signs and advertisements, which are perhaps more selling than informing. The intended readership is passers-by, drivers and potential users, that is to say anybody, essentially the ideational metafunction of language for conveying information (Halliday & Mattheisen, 2005).

Informing signs divide into two main groups. The first group consists mostly of noun phrases, and only a few finite verbs. Fig 3a is a typical opening times sign with a structured list of phrases separated by line-breaks, no punctuation except for the fromto dash and numerical stops seen in <7.30am – 5.00pm> and limited use of caps for the start of phrases/lines <Closed>. Its most prominent feature is <Opening times>, which we can call the Header, signalled partly by its greater point size and boldness. Fig 3b also has no punctuation apart from a final exclamation mark after the only finite clause and an apostrophe; line breaks mark phrase boundaries; every noun has a word-initial capital letter <Bouquets and Flower Baskets>. The sandwich blackboard on the pavement in Fig 3c starts with a Header <Coffee & Sandwich Shop.> and continues with a list of items for sale, to be discussed below. Both Figs 3a and 3c are left-aligned rather than having the central alignment of the other signs seen so far.
The estate agents’ board Fig 3d has a Header <for sale> in the middle of the sign, all in lower case. A Header is not a heading in that it can occur where the writer chooses rather than at the top of the sign. Fig 3d has an Identifier <SANDERSON YOUNG>, which is not indexically linked to the location of the sign and proclaims ownership of the sign rather than of the property. It also has two Identifiers consisting of a phone number and web address.

4a.   4b.   

Set 4. Informing signs

The second group of informing signs consists of dense texts providing information for readers on foot, mostly owned by the local authority. The job ad in Fig 4a is left aligned with two sentences, one has a short passive wanted, the other two imperative verbs call and walk-in, with an unusual hyphenation for the verb form. A more formal example is seen in the Fig 4b <Planning Notice>; by law, planning proposals have to be displayed for 21 days near the site (Development Management Procedure, 2010). Beneath a Header and an Expander, the notice is phrased in full lexical sentences; let us call this an Informer. Fig 4a therefore has only an Informer consisting of two sentences and no Header. While the details are specific to the location, most of the phraseology is laid down by the provisions of the Development Management Procedure (2010) in lexical sentences complete with full stops and initial capital letters: though owned by the local authority, planning notices are authored by them to only a limited extent. The readership are interested passersby. These are the only signs in which conventional standard punctuation is used consistently, including clause-final and linking commas. The Header in Fig 4b is then a large bold <Planning Notice> with an Expander <An application for planning permission …>. Fig 4c is similar, having a Header <paybyphone>, an Expander <the alternative way to pay> and an Informer consisting of a list of bullet-initial phrases < • no more hunting for machines…>. All of the Informing signs are indexical to their location, apart from posters or other advertisements for widely available products such as Stella Artois lager. Fig 3a announces that these premises are open at these times; Figs 3b and 3c that these goods are available at this location; Fig 3d that this building is for sale; Fig 4a that a job is available at this restaurant. Moving the signs to another location falsifies their meaning. Readership for these informing signs are passersby or drivers potentially interested in the information.
As Kress & Van Leeuwen (1996) point out, the materials that signs are made of contribute to their meaning. Locating signs are permanent and so often made of stone or metal, both for endurance and to create an impression of solidity, for example the large raised metal letters in Fig 1b <BARKER & STONEHOUSE> and the sculptured stone letters of <ASPERS CASINO>. Informing signs, however, are often ephemeral, made of paper and written by hand (Fig 3b) or PC printer (Figs 4a & 4b) or done in chalk or paint on blackboards (Fig 3c). Ephemerality may also be a matter of temporary display, whether the estate agents’ board nailed to a wall (Fig 3d) or the planning notice cable-tied to a lamp post (Fig 4b).


5a.  5b.   
5c.  5d.    5e.
Set 5. Controlling signs

Controlling signs ask or require people to behave in particular ways, whether drivers, pedestrians or customers. This may amount to control of traffic and parking, to warnings against stealing, and to suggestions how to open doors. The responsibility for these varies from traffic signs, licensed and authored by national decree but owned and erected by the local council, to warnings of wet paint put up by individual property owners. They are totally indexical in location and orientation. While Locating signs function as separators marking out boundaries, Controlling signs connect things together (Simmel, 1997).

One category is official signs controlling the movement of road-users and pedestrians according to the Traffic Signs Regulations and General Directions (2002) (encapsulated in The Highway Code (1999)), as in Figs 5a – 5d. Signs written on road surfaces have no punctuation, are all in capitals and consist of terse commands like <NO ENTRY> (Fig 5a) and <LOOK LEFT> (Fig 5b) or even simply of a picture of a bicycle (Fig 5c). Those intended for road users are written in the elongated Pavement typeface seen in Fig 5a.

The reader’s orientation to the message is part of its indexical meaning: <LOOK LEFT> only works from one side of the street and it would be catastrophic if the reader read it upside down from the other side – a classic case of Levinson’s relative direction (Levinson, 1996) as opposed to the absolute direction seen in the Toronto street sign <PEDESTRIANS CROSS AT SOUTH SIDE ONLY>. Grammatically these are mostly noun phrases, with the occasional imperative.

The type of controlling sign in Fig 5d consists of a series of punctuation-free signs (apart from numerical stops for time and fromto dashes), often starting with a capital letter <One way>. Division between phrases is by line-break except for the phrase <Goods vehicles loading only> where it is by word division. Shape and colour are significant in ways dictated by the Highway Code (1999). Traffic signs are usually made of metal for permanency.

A variation is wayfinding signs, whether owned and erected by the local council for pedestrians, as in the finger-post seen in Fig 5e <Eldon Square Shopping Centre….>, or national traffic direction schemes administered by the council. Each word has a capital letter, necessarily as these are placenames, and each noun phrase has a line-break. It is necessary for such signs to be aligned to point to the objects they refer, i.e. that the ‘text vector’, in terms of Scollon and Scollon (2003), corresponds to the direction of movement.

6a. 6b. 6c.6d.
Set 6. Controlling signs

Other controlling signs are owned by the property owner. Warnings include the no smoking sign Fig 6a, which has to be displayed in all publicly accessible buildings, the very form, words and language being imposed by law in the Smoke-free Regulations (2006). It also includes a range of signs discouraging parking like Fig 6b, and wet paint signs like Fig 6c. These lack punctuation marks and mostly have linebreaks at word divisions, apart from the no-smoking sign. Other warnings are seen in Fig 6d <THIEVES WILL BE PROSECUTED>, which seems general moral exhortation – another sign announces <Stolen Plastic cards are not welcome here>. Many warnings are all caps like Fig 6c and Fig 6d; some are word-initial caps like Fig 6b <Disabled Vehicles Only>. Few use capital letters in the same way as the standard punctuation.

7a.        7b.
Set 7. Controlling signs

Finally there are notices telling one how to press bell-pushes as in Fig 7a <Please ring the bell> and open doors as in Fig 7b <PULL>. These necessarily involve short imperatives, have no punctuation and are often all in caps. They are highly indexical in that their position has to relate appropriately to the actual objects they refer to.

Like informing signs, Controlling signs usually involve a Header, sometime by itself as in Fig 6c<WET PAINT> or Fig 7a <PULL>, sometimes a list as in <Eldon Square…> (Fig 5e). The Header may have an Expander, as in Fig 6a <NO SMOKING> and <It is against the law …>. While lists like Fig 5e are left-aligned, the remaining controlling signs are centred and symmetrical whether the writing on the road (Fig 5a) or the handwritten warning such as Fig 6c, apart from balancing an icon in the rectangle (Figs 5d & 6d).

A further characteristic of controlling signs is that they often contain small icons (Peirce, 1906), say the person pushing a trolley in Fig 5d or the one in a wheelchair in Fig 6b, the capital <M> on Fig 5c indicating a Metro station or the prohibiting circle with a diagonal line in Fig 5a, transferred from road traffic signs. Arrows are frequently present as in Figs 5d and 7b, though the sign itself may point in the right direction, as in Fig 4e. These are all iconic in that they do not simply represent an object but a meaning attached to an object, which it is up to the reader to deduce – people from outside Newcastle tend to think the <M> stands for the supermarket Morrisons rather than the Newcastle Metro.


8a  8b.  8c.8d.   8e.
Set 8. Service signs

Service signs are put up by providers of services either to tell people that the service is available at this location or to guide specialist workers to the right manhole, etc. They are owned by various specialist companies. Their readership is specialised whether to users of the service or workers for the utility service. On the one hand are general services provided for the public on foot such as telephone kiosks (Fig 8a) and pillarboxes (Fig 8b). The texts on these are minimal noun phrases – Telephone and Post Office – together with a list of collection times etc and appropriate logos such as the royal monograph <VR> (Fig 8b). The meaning is more in the iconic shape and colouring – the Stowell Street pillarbox in Fig 8b is in fact a reproduction of a celebrated 1869 Penfold design– than in the actual text.
On the other hand some signs are only intelligible to a specialist reader such as the fire hydrant sign (Fig 8c) and manhole cover (Fig 8d). These have fragments of code rather than grammatical units and lexical items of English. If you are a fire-fighter, you will know that the figure <4> above the <H> in Fig 8c gives the diameter of the pipe in inches, the figure <35> below states how far away the actual hydrant is in feet. But who knows what lies beneath the manhole cover labelled <CATV> (Fig 8d)? Apparently the easy deduction that it is cable television does not work as these covers are used by several services. Hence the addressed readership is not the ordinary passerby or driver but a service worker.

The readership for the ubiquitous burglar alarm signs such as <ADT> (Fig 8e) is more problematic. Partly they announce indexically that these premises are fitted with alarms, helpful for insurance purposes and to deter burglars; partly they advertise a service. Grammar and punctuation are more or less irrelevant to service signs as they do not usually have grammatical structures.

Other types of signs not dealt with here include monumental signs such as inscriptions on war memorials, graffiti and artworks such as the Invader signs by a French street artist dotted around Newcastle. These genres need separate treatment.


In discussing the language of the street, Cook (2013), following Leech (1966), distinguishes block grammar that lacks articles etc from abbreviated grammar which has a wider range of constructions, such as the imperatives Please ring the bell (Fig 7a). Locating signs and service signs necessarily consist mostly of a proper name like Greggs. Locating signs always use block grammar for Locators consisting of noun phrases without article as in King Neptune (Fig 2a) and Expanders as in kingswalk dental implant practice (Fig 2b). Hence the presence or absence of an initial capital letter is slight evidence for a word in a Locator being a proper noun. The two groups of informing signs have different grammars. One group use block noun phrases for Headings and Lists such as Opening times (Fig 3a) and Savouries & Cakes (Fig 3c). The other group have ‘full’ lexical sentences, whether for jobs (Fig 4a) or for planning (Fig 4b) Please walk-in to apply …. Controlling signs mix noun phrases such as WET PAINT (Fig 6c) with full sentences with imperatives Please ring the bell (Fig 7a) and impersonal prohibitions with full verbs It is against the law to smoke in these premises (Fig 6a). Service signs either have noun phrase Headers Telephone (Fig 8a) or cryptic messages which can barely be called phrases CATV INTEGRAL SLIDEOUT B125 (Fig 8d).

The grammar of the street is highly restricted, mostly consisting of block grammar noun phrases. The occasional lexical sentence is confined to informing signs. What seems crucial is position and prominence rather than linear order. According to Kress and van Leeuwen (1996, p.223), ‘Non-linear texts... select the elements that can be viewed and present them according to a certain paradigmatic logic, the logic of Centre and Margin or of Given and New, for instance, but leave it to the reader to sequence and connect them’. It must not, however be forgotten that such statements are not as yet backed by objective evidence, such as the use of eye-tracking. Vertical arrangement in lines is undoubtedly crucial, though not necessarily read from top to bottom of the sign. The distinctive grammar of street signs puts rather different demands on the punctuation system. It is marking out nouns and noun phrases that is important rather than distinguishing clauses and sentences.


9a.   9b.
Set 9. Lists

We can now return to the List structure, which can occur within any of the types of signs. Fig 1d provided a typical List <Cask Conditioned Ales …> – a series of equivalent items detailing what’s on offer, consisting of noun phrases marked out by a change of colour with each noun having initial caps apart from <spirits>. The division between list items is a line-break with additional leading. Fig 9a shows a more conventional use of commas as a list separator <2, 3, 4, 5 & 6>; Fig 3c uses full stops for the same effect <Jacket Potatoes. Salads. Savouries and Cakes.>; Fig 4c shows a list separated by bullets <• no more pocketfuls of change>, under the influence of word processing. The list structure in Fig 9b depends on layout in two columns <Americano Mocha>. In most Lists like Figs 1c and 3c, all the nouns have word-initial capitals and the division between items is through line-breaks.


We can now compare the overall use of punctuation marks in street signs with the standard punctuation in written English texts. Table 2 gives the frequencies of standard English punctuation marks calculated in three ways. [Some figures calculated after the apperance of this paper are given in Frequencies for English Punctuation marks]

Table 2. Average occurrence per 1000 words

The first column is based on a writing system corpus some 459 thousand words long. This includes three novels of different types (276 thousand words), selections of articles from two newspapers (55 thousand), one bureaucratic report (94 thousand), and assorted academic papers on language topics (34 thousand). More details are provided on The figures are presented as an average per thousand words of text.

The second column is the Ngram count for the year 2000 British and American Google corpus, again converted to scores out of a thousand words. It does not include commas as these serve as a delimiter in Ngram Viewer and so cannot be counted. The third column is the count for the Corpus of Contemporary American English (2013) based on 450 million words. In this case the double quotation mark was not searchable as it appears to be stripped from the search item. A more delicate frequency analysis of commas in the Wall Street Journal by Bayraktar et al (1998) showed most commas separated ‘appositives’, items in lists and sentence initial phrases.

It is surprising that there are such differences in the averages between the frequencies for the three samples, given the large size of the Ngram and COCA corpora. The explanations for this we will not develop here, quite possibly attributable to the different ways in which frequency counts for punctuation are made in the three corpora.

The two most frequent marks are full stops and commas. While the different non-grammatical uses of full stops and commas, for example abbreviations and numerical stops, cannot be separated out, clearly these are the major way of showing grammatical units. Next in frequency come double and single quotation marks, again qualified by the inability to tell which of their uses is intended, say relational apostrophe versus closing quotation for < ’ >. The overall average for our corpus is 235.2 punctuation marks, or 1 for every 4.3 words.

The language of the street, however, has far fewer grammatical punctuation marks. The corpus used here numbered 296 signs. Of these 182 had no punctuation marks other than apostrophes and numerical stops, i.e. 61.5%. Excluding fromto dashes brought the total of punctuation-free signs up to 196, i.e. 66.2%. 26 signs (8.7%) made some use of full stops; excluding numerical stops and abbreviations, the total came to 12 (4.1%).
Calculating frequencies per 1000 words is hard for our sign corpus since some signs like Fig 4c have no written text. The total word count came to 3412 words. The total of full stops was 127, yielding an average of 37 per 1000 words, considerably down on the 235.2 for standard texts given in Table 2. Setting aside numerical stops, the total was 41, i.e. an average of 12 per 1000 words. Commas occurred 39 times, or 11.43 times per 1000 words, compared to the standard 64.1 times. 18 of these were listing, i.e. 46% compared to the 20% found by Bayraktar et al (1998).

The language of the street thus uses punctuation very sparingly. How then can it convey the necessary grammatical information? To answer this, let us try to capture the punctuation of the street in a similar fashion to the presentation of standard punctuation in Table 1.

Going down the grammatical ranks:

- marking of sentences is not needed. Only informing signs use lexical sentences finishing with a full stop and starting with a capital letter. Occasional exclamation marks are found in informing signs with an imperative <BOOK NOW!>; there are no question-marks on any signs. Hence initial capital letters and final punctuation marks are seldom a clue to the beginning of a sentence, caps being used in other ways.

- marking of clauses by commas etc is absent even from informing signs like Fig 4c <If you would like to discuss the matter you can phone the officer …>.

- marking of phrases is mostly achieved through line-breaks, both between phrases as in <The Gate [line-break] Grey’s Monument> (Fig 5e) and within phrases as in <Fine [line-break] Wines and spirits> (Fig 1d). As Waller (1990) points out, line-breaks by themselves may be insufficient to be meaningful as they could also be purely arbitrary divisions of text as in Fig 2b. So line-breaks can be reinforced by bullet points (Fig 4c) or by change of colour (Fig 1d). One phrase finishes with an exclamation mark <This weekend only!>. Phrases vary between all capitals <LEAZES PARK RD.> (Fig 1c) and all lowercase <percy house> (Fig 2d); a few have phrase-initial caps <Opening times> (Fig 2a).

- marking of words is by space or line-break. Some have word-initial capital letters <Disabled Vehicles Only> (Fig 6b) and <Available from this Summer>. Unlike standard punctuation, an initial capital letter does not indicate a proper noun, except perhaps in Informing signs. Take for example Fig 9d with its striking lowercase <t> dotted between the capital letters <StUDENt APARtMENts>.
- marking of morphemes is through possessive ’s. It is nevertheless frequently omitted, as in <Lisas Coffee Shop>.

- letters. The full stop is sometimes used for abbreviations as in <LEAZES PARK RD.> but not always – the other end of the street is signposted <LEAZES PARK ROAD> – and in <MR. M.F. LEE BDS>. Another angle is the substitution of punctuation marks for letters in propernames of shops, which does not occur in our sample but is widespread in shopnames like <sk:n> and <b:zarre> etc, which opens up the whole issue of novel invented spellings that will not be included here. Further examples can be found on XXXXXX. The only parallel example is the replacement of <I> by a tooth implant in Fig 2b.

We can now reverse the approach and see how the categories of the structure of signs we have postulated are marked out in various ways.

- Locators state the identity of the property by number (Fig 1a) or name (Fig 1b), hence they typically consist of noun phrases. The punctuation is division by word space and line-breaks. Compared to other uses, the letters tend to be either all caps (Fig 1b) or all lowercase (Fig 2d).

- Headers give the main point of the sign, usually a noun phrase (Fig 3c), mostly without punctuation other than line-breaks (Fig 1d). They are similar to Locators in standing out from the rest of the sign in prominence but may be a variety of phrases (Figs 3d & 5d).

- Expanders can be different types of phrase, whether verbless clause (Fig 1d) or noun phrases (Fig 2b). Typically they have no punctuation other than line-breaks and are less prominent than the Header or Locator.

- Identifiers give additional information such as addresses, web addresses and phone numbers, using address punctuation of commas and numerical dots (Fig 2c). These may have no punctuation (Fig 2c)), and often have least prominence in appearing at the bottom in smaller font sizes (Fig 3d).

- Informers provide detailed information, mostly through full lexical sentences with initial caps and final full stops (Figs 4a & 4b), sometimes bulleted lists (Fig 4c).

- Lists are structures of similar items, punctuated by line-breaks, commas etc (Figs 9a & 9b) and by layout and by change of colour (Fig 1d).


The general paper on the language of the street (Cook, 2013) raised a number of issues about the language roles related to signs, the varieties of street sign and the distinctive nature of their language. The present paper has developed punctuation as one aspect of this complex linguistic situation and has shown how punctuation functions differently in this genre from the ‘standard’ variety, linked to its different grammar. Punctuation adapts itself to different circumstances and uses new markers such as line-breaks. The punctuation of the street is not deviant, illiterate or misguided; it is a response to the functional needs of those who write and read it.


Austin, J. L. (1962). How to do Things with Words. Oxford: Clarendon Press.
Bayraktar, M., Say, B. & Akman, V. (1998). An analysis of English punctuation: the special case of the comma. International Journal of Corpus Linguistics 3, 1, 33-57.
Ben-Rafael, E., Shohamy, E., Amara, Muhammad, H. & Trumper-Hecht, N. (2006). Linguistic landscape as symbolic construction of the public space: the case of Israel. In Gorter, D. (ed.) Linguistic Landscape: a New Approach to Multilingualism. Clevedon: Multilingual Matters. Gorter (ed.), 7–30.
Biber, D., Johansson, S., Leech, G., Conrad, S. & Finegan, E. (1999). Longman Grammar of Spoken and Written English. Harlow: Longman.
Bloomfield, L. (1933). Language. New York: Holt.
Carey, G. (1960). Mind the Stop. Cambridge: Cambridge University Press.
Chomsky, N. (1995). The Minimalist Program. Cambridge, Mass.: MIT Press.
Cook, V. J. (2004). The English Writing System. London: Hodder.
Cook, V. J. (2010). 'Prolegomena to second language learning', in P. Seedhouse, S. Walsh & C. Jenks (eds.), Conceptualising Language Learning, Palgrave MacMillan, 2010
Cook, V. J. (2013). The Language of the Street. Applied Linguistics Review, 4, (1), 43 – 81.
Companies (Trading Disclosures) Regulations, The (2008).
Corpus of Contemporary American English (2013).
Development management procedure (England) order 2010 (SI 2184) (2010).
Goffman, E. (1963). Behavior in Public Spaces. New York: Free Press.
Hall, S. (2012). City, Street and Citizen: The Measure of the Ordinary. Routledge
Halliday, M. A. K. (1985). Spoken and Written Language. Oxford University Press.
Halliday, M. A. K. & Mattheisen, C. (2005). An Introduction to Functional Grammar. London: Hodder.
Highway Code, The (1999). London: the Stationery Office.
Jacobs, J. (1961). The Death and Life of Great American Cities. New York: Random House.
Kress, G. & van Leeuwen, T. (1996). Reading Images: the Grammar of Visual Design. London: Routledge.
Leech, G. N. (1966). English in Advertising. London: Longman.
Leech, G. N. (2013).
Levinson, S. C. (1996). Relativity in spatial conception and description. In J. J. Gumperz & S. C. Levinson (eds.), Rethinking Linguistic Relativity. Cambridge: Cambridge University Press, 177-202.
Lowth, R. (1775). Short Introduction to English Grammar. Delmar, N.Y.: Scholars' Facsimiles & Reprints, 1979.
McCaskill, M. K. (1998). Grammar, Punctuation, and Capitalization: a Handbook for Technical Writers and Editors. Hampton, Virginia: NASA SP-7084, on-line at
Nunberg, G. (1990). Linguistics of Punctuation. Stanford: Center for the Study of Language and Information.
Outdoor advertisements and signs: a guide for advertisers (2007). London: Department for Communities and Local Government.

Parkes, M.B. (1992). Pause and Effect. Burlington, VT:Ashgate
Peirce, C. S. (1906). Prolegomena to an apology for pragmaticism. The Monist, XVI. 4. 492–546.
Puttenham, G. (1589). The Arte of English Poesie. London: Richard Field.
Scollon, R. & Scollon, S. (2003). Discourses in Place: Language in the Material World. London: Routledge.
Simmel, G. (1997). The sociology of space. In Frisby, D. & Featherstone, M. (eds.), Simmel on Culture, London: Sage Publications, pp. 138-170.
Steinhauer, K. (2003). Electrophysiological correlates of prosody and punctuation. Brain and Language, 86, 142–164.
Smoke-free (premises and enforcement) regulations, The (2006).
Todd, L. (1995). The Cassell Guide to Punctuation. London: Cassell.
Traffic Signs Regulations and General Directions (2002). UK Government,
Trask, L. (1997). The Penguin Guide to Punctuation. London: Penguin.
Venturi, R., Brown, D. S. & Izenour, S. (1977). Learning from Las Vegas. Revised Edition. Cambridge, Mass: MIT Press. First published 1972.
Vienna Convention on Road Signs and Signals (1968). UNESCO.
Waller, R.H.W (1990). ‘Typography and discourse’, in R Barr et al (eds), Handbook of Reading Research, vol II, New York: Longman, 341-380.
Wilson, P. (2007). Punctuation Elementary Guidance.