The choice of directive expressions in English and Spanish

instructions:
a semantic network
Susana Murcia-Bielsa
University of Stirling

Introduction

In this paper I examine the range of expressions in English and Spanish consumer
product instructions that are commonly known as ‘directives’: expressions whose aim is
to get the addressee to perform an action or a group of actions, and whose typical
realisation is thought to be the imperative form. The imperative, however, is not the
only directive form; there are many other directive expressions available. This study,
based on a corpus of naturally-occurring instructions for consumer products in English
and Spanish, sets out to find the reasons behind the existence of the wide range of
directive realisations. Drawing upon Biber’s (1995) register analysis framework, it will
be shown that situational features are related to linguistic forms through functional and
conventional associations. These will be presented through two semantic networks (one
for English and another for Spanish) showing the factors influencing the choice of
expression. With these networks I aim (i) to explain the meanings conveyed by each
realisation and (ii) to expose the similarities and differences between Spanish and
English directives in instructional texts.
This paper will start by explaining what a directive is, what classifications of
directives have been offered in the literature, and where - within those classifications -
directives in instructional texts fit. A brief overview of the data used will then be
provided. The remaining sections of this paper will be structured following Biber’s
framework for register1 analysis.
According to Biber (1995: 10), there are three major components in register
studies: “description of the situation in which the register is used; description of the
linguistic characteristics of the register; and analysis of the functional or conventional
associations between the situational and linguistic features.” He schematises these
relationships as in Figure 1. The relationships are bi-directional, i.e., situational
characteristics influence the choice of linguistic form, and at the same time the choice of
linguistic features helps create the situation.2
INSERT FIGURE 1 HERE
Followig this pattern, the results of my analysis will be presented as follows: firstly, I
will focus on the linguistic forms and present the directive expressions found in the
data. I will then show that situational features explain the appropriateness of the most
frequent realisations found in the data. The following section focuses on the
associations between the situation and the form, showing the factors that influence the
choice of expression and formalising them in a semantic network that can be a useful
tool for text generation. Finally, the results for both languages will be compared,
establishing similarities and differences between Spanish and English.

Directives in instructions

1
Before proceeding to study the particular realisations, it is important to specify what is
meant here by ‘directives.’ This section will draw upon the literature in order to provide
a starting point for a characterisation of directives in instructions.
Directives have been defined in Speech Act Theory (Austin 1962; Searle 1969).
The notion of speech act is briefly summarised by Crystal (1992: 362) as “a
communicative activity defined with reference to the intentions of a speaker while
speaking and the effects achieved on a listener”. It is precisely the incorporation of the
intentional aspect that makes the notion of speech act especially appealing here; as
noted by Paris and Scott (1994), intentions play an important role in instructional texts
and the directive stance or intention is one of the main features of instructions.
Directives are one of the five basic kinds of speech acts proposed by Searle
(1976), and are defined as “attempts ... by the speaker to get the hearer to do
something” (Searle 1976: 11). They can be conveyed in many different ways.
According to Huddleston (1984: 351), for instance, a directive “is a term that covers
requests, commands, prohibitions, instructions and the like.” Longer lists of directive
speech acts have been suggested by Searle (1976, 1979); however these lists are not
particularly relevant for my purposes. I will rather adopt the classifications in Ross
(1968) and Ervin-Tripp (1976) as these take into account the situational features.
Ross’s classification is broader in the sense that it helps to provide a
characterisation of directive texts rather than individual directive expressions.
Therefore, it is relevant here as a starting point for setting the general features of
directives in this particular genre. Ross (1968) distinguishes among different types of
directives according to differences in situation and motivation.3 The directives found in
the instructional genre belong to Ross’s category of personal directives and more
specifically to the hearer-interested type. According to Ross (1968: 38), a personal
directive has a clearly defined sender (A) and recipient (B). In the register under study
here, A represents the manufacturer and B represents the consumer. Even though these
two participants are not always explicitly signalled in the data they can be retrieved
from the context, therefore they are clearly identified. The aim of hearer-interested
directives is “to produce actions which it is in B’s interest to perform.” According to
Ross, the linguistic-contextual manifestations of this directive category are advice,
warning, recommendation, and direction. As will be shown below, these are all found
in the data.
Ross’s classification is useful because it allows us to differentiate instructions
from other directive texts (whether they are spoken or written). It also differentiates
consumer product instructions from other instructional texts such as rules for games, for
example. The latter are impersonal directives which have no sender and recipient as
such, but are constitutive rules based on mutual agreement.
The second useful classification links contextual features to types of realisations.
Although Ervin-Tripp’s (1976) classification was intended for spoken English, it is also
applicable to written directives. She suggests the following six kinds of directives,
ordered approximately according to the relative power of speaker and addressee in
conventional usage and the obviousness of the directive (Ervin-Tripp 1976: 29):

Need statements, such as I need a match.

Imperatives, such as Gimme a match and elliptical forms like a match.
Imbedded imperatives, such as Could you gimme a match?

2
 Permission directives, such as May I have a match?
Question directives, like Gotta match?
Hints, such as The matches are all gone.

Not all the types in Ervin-Tripp’s classification of directives occur in written

instructional texts: in my data there are no permission directives, and imbedded
imperatives are extremely rare.4 However, it is possible to find cases of the remaining
four types of directives, imperatives5 being by far the most frequent, as will be shown
later. Very few cases of need statements, question directives and hints were found.
The differences between the forms seem to lie in the fact that different degrees of reader
inference are needed in each case. Need statements such as (1) below express a
necessity and, consequently, imply an obligatory action on the part of the reader.
Question directives such as (2) are exclusively found in the troubleshooting section,
and are obvious directives in the instructional genre. Hints such as (3), which are much
less frequent, require a greater inference chain; they are more indirect and require
considerable knowledge of the situation.
(1) The authorisation code number will need to be stored into the Mercury memory
location. [Morphy Richards telephone] 6
(2) Does the paper bag need emptying or replacing? [Hoover vacuum cleaner]
(3)The use of abrasive scouring pads or cleansing powders can mark the surface.
[Prestige pressure cooker]

Taking into account the definitions and classifications provided in the literature, I will
now summarise what is meant here by ‘directive.’
• The aim of directives in instructions is to get the reader to do (or not to do) the
actions in the task plan, the task plan being a hierarchical representation of the actions
for performing a task (Sacerdoti 1977). Negative directives have been counted together
with the ones with positive polarity in the tables showing the frequency. The networks,
however, do not reflect negative polarity.7
• Since the purpose of instructional text is to get the reader to use the product “safely,
efficiently and correctly” (Delin et al. 1993; Paris and Scott 1994), the directives in this
genre are clearly hearer-interested (Ross 1968). In other words, the actions are of
benefit to the addressee.
• Following Ross (1968), the linguistic manifestation of the directives in this genre
covers direction, warning, advice and recommendation.
• Indirect directives such as those classified by Ervin-Tripp as question directives and
hints have in general been excluded from this study. Even though the context provides
all the elements necessary for their interpretation as directives, it was thought more
appropriate to leave them out from the networks and focus only on ‘direct’ directives.
Only one exception to this was made: hints such as (4) - realised by the present tense in
the passive voice - were included in the data analysis. The reason for their inclusion is
related to the comparison between Spanish and English: Spanish has an impersonal
present tense in the passive voice which is unmistakably recognised as a directive.

(4) ... optimum performance is achieved by preheating the grill for about 1 minute.
[Creda cooker]

3
Having made clear what is meant in this paper by ‘directives’, the next section will
provide a brief introduction to the data and how it was analysed.

The data and method of analysis

The study was performed on the basis of two separate corpora (differentiated according
to language) of instructions for household and consumer appliances. Both corpora
contain only instructions that originated in the language of study, rather than
translations; translations present the danger of being influenced by the source language
on the discourse and the grammatical levels, which might lead to unreliable results. The
data, then, are based on ‘paired texts’, i.e., parallel texts that Hartmann (1980, 1996)
defines as texts that “are not translationally equivalent, but functionally similar in
situational motivation and rhetorical structure ... from any pair or multiple of
languages” (1996: 950). The English data consist of 1087 directive expressions
extracted from 10 different user instructions / manuals for household appliances, a total
of 25,897 words. The length of the texts ranges from just 184 words to over 7,000. The
Spanish data consist of 716 directive expressions extracted from 17 different sets of
instructions, a total of 21,552 words. The average length of the Spanish texts was
shorter than that of the English texts, but the range of text lengths is still wide: from 400
words up to 3,711.
The extraction of the directives was done manually, by selecting all the tokens
which prompted an action from the reader in a direct way, i.e., without the need for a
long inference chain. The tokens were then divided according to the linguistic
expression they contained. The expressions of directive found in the data will be
presented in the following section.

Linguistic forms: directive expressions in the data

The imperative clause is generally thought to be the syntactic category that typically
realises a directive. It is therefore possible to assume that each time a directive is
intended, an imperative will be used, as in (5)8 and (6).
(5)Descuelgue el microteléfono. [Telefónica telephone]
(Lift the handset.)
Lift the handset. [Morphy Richards telephone]

INSERT TABLE 1 HERE

INSERT TABLE 2 HERE

Tables 1 and 2 show that the imperative is indeed the most frequent form used for
realising directive intention in instructional texts. However, a wide range of other
expressions is also used. My analysis showed the following directive realisations - in
addition to the imperative - which are available in English instructions:
Modal: declarative clauses with the modals must (as in (7)) and should (as in (8)).

(7) This Instruction Book must be kept handy for reference. [Hotpoint refrigerator]

4
(8) The bowl should be washed and dried after use and before storage. [Boots kitchen
scale]

Appeal to reader: expressions such as “it is + adjective/participle + action” and “we

recommend + action”, which appeal to the reader to act by presenting the benefits of
performing the action (as in (9)).

(9) It is a good idea to clean the inside of your fridge after defrosting. [Hotpoint
refrigerator]

Other: Non-modalised expressions such as the future tense, as in (10), or the present
tense in the passive voice, as in (11).

(10) The minimum height of the cooker will be set at 900mm to the top of the hob. [Creda
cooker]

(11) ... optimum performance is achieved by preheating the grill for about 1 minute.
[Creda cooker]

In turn, the following expressions are available for Spanish directives in addition to the
imperative:
- The Infinitive,9 as in (12),

(12) Limpiar con una esponja o estropajo metálico, utilizando detergentes o jabones ricos
en sosa. [Corberó cooker]
(Clean with a sponge or metallic scourer, using detergent or soda-rich soap.)

- Modal: declarative clauses with the modal deber, as in (13).

(13) Siempre debe cubrir los alimentos a freír. [Jata deep fryer]
(You must always cover the food when frying.)

- Appeal to reader: expressions which appeal to the reader to act, such as (14).

(14) Es conveniente limpiar periódicamente los casquetes y quemadores. [Corberó

cooker]
(It is convenient to clean frequently the caps and burners.)

-The Future tense, as in (15).

(15) Una vez conectado el aparato, se seleccionará con el interruptor la temperatura de

aire deseada. [Ufesa fan heater]
(Once the device is connected, the required temperature will be selected with the knob.)

- ‘Se’ present: present tense in the passive voice with ‘se’, as in (16).

(16) Se abre el Grill y se coloca el alimento a asar. [Magefesa grill]

(It-is-opened the grill and it-is-placed the food to-be-cooked.)10

5
When looking at the realisations and frequency tables for Spanish and English, some
differences between both languages are noticeable at first sight:
• Spanish has a realisation that is not available in English: the infinitive.
• In the data, English can express a directive through two modalised realisations (must
and should), while Spanish has only one modalised realisation available (deber), despite
its modal system also having a more tentative (or weak) form expressed through the
conditional tense (debería), which is mainly used in speech for giving advice or making
suggestions.
• The data show that both languages can express direction through the present tense in
the passive voice. However, there is a slight difference between Spanish and English:
this realisation exclusively signals directive intention in Spanish, while it can express
either informative or directive intention in English.
• The frequency tables (Tables 1 and 2) show that the imperative is much more
frequent in English (79.30%) than it is in Spanish (52.93%). The difference between
Spanish and English imperative frequencies might be due to the existence of the
infinitive as a directive - a form unavailable in English. When taken together, the
Spanish imperative and infinitive data total 74.16%, comparable to the English
imperative level of 79.30%.
• The frequency of occurrence of the modalised realisations is similar in both
languages when the results for the two English modals are counted together: 12.43% of
modalised directives for Spanish and an overall frequency of 14.72% for English.
This section has shown the wide range of forms available for expressing
directives both in English and in Spanish instructions. The following sections will set
out to explain the appropriateness for instructions of certain forms which would be
considered inappropriate in other contexts, as well as the reason for the wide variety of
expressions available.

Situational features: appropriateness of certain expressions

Directives are inherently face-threatening (see Brown and Levinson 1987), since they
impose a demand for behaviour on the addressee. There are different ways of lessening
this threat: in speech, for instance, this is done through the use of requests, which leave
the addressee the option not to act and are usually realised by particular modal verbs
(e.g., could, would, might). My data, however, rarely exhibit any lessening of that
‘inherent’ threat. As the frequency tables show, the imperative - a form frequently
considered inappropriate in most contexts for politeness reasons - is by far the most
frequent realisation of a directive in instructions in both languages; a high frequency of
occurrence is also exhibited by modalised directives with must and should in English
and deber in Spanish.
Butler (1988), in his study of the modalised realisations of directives in English,
supports the hypothesis that politeness greatly motivates the choice of expression. He
claims that, in speech, straight imperatives and modalised statements with will and must
(expressing the imposition of an obligation), leave no option apart from outright refusal
to comply and are therefore classified as orders, which tend to be impolite. My data,
however, suggest that at least in the case of instructional texts politeness does not play
such an important role. Butler’s (1988) suggestion that the imperative and the

6
modalised directives with must are impolite does not hold for instructional texts, where
the beneficiary of the act is the addressee, as discussed in the next subsection.
This section on situational features sets out to explain the appropriateness for
instructions of those directive expressions that are most frequently considered impolite
or inappropriate in other contexts. In order to do this, we need to look at what
influences the choice of expression. The following quotation from Biber suggests some
factors influencing linguistic form:

Associations between form and situation can be motivated either by

functional communicative requirements or by simple conventions.
Functionally motivated patterns can be related to a number of situational
characteristics such as the physical setting, the extent of shared context or
background knowledge, the degree of interactiveness, the production
circumstances, the primary purposes or communicative goals of
participants, and the social relations among participants. (Biber 1995: 10)

It will be argued in the following that syntactic expressions do not have a fixed
politeness level; their degree of politeness, rather, depends on contextual features such
as those mentioned in Biber (1995). These features, which in addition characterise the
genre of instructions, will be grouped under the following headings:
• Directives in speech and writing: to refer to issues of the production circumstances
and the physical setting.
• Authority relations: dealing with shared context or background knowledge, the
degree of interactiveness, and the social relations among participants.
• Task-relevance: related to the primary purposes or communicative goals of
participants.

Directives in speech and writing

Directives are interpreted differently depending on the mode of communication, that is,
whether they appear in speech or in writing. Directives which, for instance, are polite
and taken as advice in a spoken situation would be interpreted in written instructions as
strong directives, i.e., as necessary actions. Let us look at the sentence in (17):

(17) You must look at these photographs, they're brilliant!

In spoken discourse between two women who are friends, for example, (17) would be
interpreted as a piece of advice, and the hearer would have the option to follow it or not.
However, in written instructions the same realisation would be interpreted as a
necessary action and the reader would normally do as told without any questioning, as
in (18).

(18) You must read these instructions prior to using the appliance and retain them for
future use. [Creda cooker]

In addition, the imperative, which is usually perceived as rude in speech, can be

perfectly appropriate in written instructions. Observe (19):

7
(19) Fill in the form and send it to your local office.

Between two students, (19) with no further explanation would sound rude. However,
(19) might also be appropriate in a similar speech situation between two students if they
both share the knowledge that tomorrow is the deadline for applying for a grant. It is
then possible to suggest that the imperative is polite if, and only if, the action it
expresses is of any benefit to the addressee. This has already been noted, for instance,
by Downing and Locke (1992: 198):

The more the action is likely to benefit the addressee, the more socially
acceptable an imperative will be. Otherwise, an imperative is likely to
sound curt or demanding in English.

The sentence in (19) would be perfectly appropriate in a written instruction manual. In

the context of instructional texts, performing the actions in the task plan will be of
benefit to the instructee, therefore the imperative can be used.

Authority relations

The relations between the participants are particularly important, as mentioned by Sager
et al. (1980: 27):

... in all directive speech acts there is an unequal relationship in that the
speaker is in some way superior to, in a position to direct, order, or
command the listener, no matter to what extent, tact or social practice
modifies the relationship.

In instructions the writer adopts a position of authority which is acknowledged by the

reader. This authority is not given by social status but by the instructor’s superior
knowledge about the product. The writer instructs a hypothetical reader who does not
know how to use the product properly. It is precisely this instructor’s superior
knowledge that makes it possible to use straight directives and modalised directives
with must which are considered impolite in most situations (see Butler 1988).
Despite Butler’s (1988) hypothesis that directives classified as orders will be
relatively impolite, the imperative in instructions is not considered impolite because,
rather than the writer leaving no option and imposing her11 authority, there is no other
option if the product is to be used correctly and without problems. Sager et al. (1980:
164) claim that the recipient considers instructions as “necessary guides for the
performance of his work.” The directions are given for the benefit of the reader and are
imposed by the circumstances, rather than by an individual’s will, therefore they do not
call for negative politeness strategies. In other words, they do not need any face-threat
lessening strategies.
Therefore, an explanation of the appropriateness of the imperative and its high
frequency in my data may be found in the fact that, despite authority relations,
instructions are far from being perceived by the reader as face-threatening orders. They
are rather considered as advice or guides (see Sager et al. 1980: 164).

8
 It emerges from this that the power relations involved in written instructions do
not have to do with rank or social status - as suggested, for instance, by Ervin-Tripp
(1976) for spoken directives - but, rather, with the authority coming from the
instructor’s superior knowledge, which is acknowledged by the reader because he
knows that, as discussed in the previous subsection on directives in speech and writing,
the instructions are of benefit to the instructee.

Task-relevance

A further reason for the high frequency of imperatives can be found in Ervin-Tripp
(1976: 59):

Task-relevant and role-appropriate directives are more likely to be direct rather

than imbedded or embroidered imperatives since they need to be marked less for
attentional purposes.

In other words, directive texts such as instruction manuals, whose priority is to get the
reader to perform the task properly (i.e., texts that are task-relevant), are most likely to
avoid the use of embroidered and indirect imperatives such as imbedded imperatives or
permission directives, since they would draw the reader’s attention away from the task-
plan (see Dixon 1982, 1987a, 1987b; Dixon et al. 1988; and Agre and Batali 1991, for
further research on task-relevance issues).
The corpus used in the current study is unquestionably task-relevant. As often
stressed by the kind of message found in instruction manual guarantees (as in (20)
below), it is necessary to follow the instructions provided in the manual. The
instructions accompanying any product are the key to making the most out of it, and
therefore it is important that the actions are clearly stated. This is best achieved by
using a direct style that draws the reader’s attention to the actions in the task plan, rather
than by using the sophisticated expressions whose main aim is to lessen the ‘threat’
posed by directives.

(20) To qualify for the 2 year guarantee the appliance must have been used according to
the manufacturer’s instructions. [Morphy Richards coffee-maker]

To sum up this section, the comparison of directives in speech and writing shows that
syntactic expressions do not have a fixed politeness level. The degree of politeness and
appropriateness of certain realisations (e.g., the imperative), rather depends on content
and context features such as authority relations, benefit to addressee and task-relevance.
If genre is defined by what Hasan (1978: 231) calls its ‘contextual
configuration’, that is, the combination of field, tenor and mode, the features discussed
in this section can be said to be characteristic of the genre of instructions for the
following reasons:
• The issues referring to the production circumstances and the physical setting
discussed in the subsection on directives in speech and writing are related to mode.
• The subsection on authority relations dealt with shared context or background
knowledge, the degree of interactiveness, and the social relations among participants, all
of which can be ascribed to the tenor.

9
• Finally, the issues discussed in the subsection on task-relevance, that is, the primary
purposes or communicative goals of the participants (the task plan), correspond to field.
Having discussed the features that can be considered characteristic of the genre
of instructions, the following section will focus on the factors influencing the choice
between the wide range of directive expressions available for instructional texts in
Spanish and English.

Factors influencing the choice of directive expression

It will be argued in this section that the wide range of directive realisations that are
found in instructions is due to the fact that each expression conveys different pragmatic
or semantic meanings. Or rather, to use Biber’s terminology, the expressions under
study here have different functional features (i.e., they perform different discourse tasks
and reflect aspects of the communicative situation and production circumstances). This
endorses, once more, the close relationship between context and linguistic form. Later
on in this section these relationships will be represented more clearly in two semantic
networks for choice of directives.

Action-relevance

One of the factors that clearly influences the choice of directive expression is what I
will call action-relevance. This factor, which is related to the situational features and in
particular to the task-plan, refers to the fact that not all the actions in the task-plan are
equally important; some actions are strictly necessary to get the device to work, while
others simply help to achieve better performance.
I will therefore argue that the actions in the task plan have different relevance for
the goal of getting the device to work and can be represented according to a cline
(Figure 2) that includes the categories of necessity and desirability in their positive and
negative polarities, thus rendering the following four types of action: necessary actions,
desirable actions, undesirable actions and prohibited actions.
INSERT FIGURE 2 HERE

The top end of the cline in Figure 2 corresponds to those actions with the highest
relevance in the task plan (namely, necessary actions). These actions are strictly
necessary to achieve the purpose stated beforehand and therefore present no options at
all. One clear example of necessary action is ‘connecting the device to the mains
supply’; without this action the device would not work.
It was mentioned above that the obligation to do the actions is imposed by the
circumstance of trying to get the device to work. It is therefore possible to assume that,
by default, all the actions are necessary or obligatory, unless otherwise stated. This
default obligatoriness, which I will refer to as unmarked obligation, is realised by the
imperative forms in both languages (as in (21) and (22)) as well as by the Spanish
infinitive (23), se passive (24) and future (25).

(21) Lift the handset. [Morphy Richards telephone]

10
(22) Descuelgue el microteléfono. [Telefónica telephone]
(Lift the handset.)

(23) Limpiar con una esponja o estropajo metálico, utilizando detergentes o jabones ricos
en sosa. [Corberó cooker]
(Clean with a sponge or metallic scourer, using detergent or soda-rich soap.)

(24) Se abre el Grill y se coloca el alimento a asar. [Magefesa grill]

(It-is-opened the grill and it-is-placed the food to cook.)12

(25) Una vez conectado el aparato, se seleccionará con el interruptor la temperatura de

aire deseada. [Ufesa fan heater]
(Once the device is connected, the required temperature will be selected with the knob.)

Despite the default urgency of the actions in the task plan, necessity marking implies a
stronger requirement for an action which might otherwise be overlooked by the reader.
The second category in Figure 2 represents those actions in the task plan which
are not totally necessary, but present some advantages for the performance of the task
and are therefore desirable. Desirable actions have the following effects: on the one
hand, the action is presented as optional, leaving its performance up to the reader; on the
other hand, though desirable actions do not claim a strong obligation, they are presented
as positive and advantageous actions for the correct performance of the task.
Consequently a positive response from the reader is most likely to occur, in view of the
advantages that performing the action offers (whether these benefits are specified or
not).
As for the linguistic realisations of any of the desirability choices, these are
always expressions of appeal to the reader, both in English and in Spanish:

(26) It is a good idea to clean the inside of your fridge after defrosting. [Hotpoint
refrigerator]

(27) Es conveniente limpiar periódicamente los casquetes y quemadores. [Corberó

cooker]
(It is convenient to clean frequently the caps and burners.)

Taking all this into account, the linguistic forms realising the different degrees of
necessity (or action-relevance) of the actions in the task plan can be represented along a
cline as shown in Figure 3. This necessity cline presents some similarities to the one
suggested by Halliday (1985: 357) for modulation (of the obligation type). For
instance, if divided into two parts, each part in my necessity cline corresponds to a
different kind of Value (High and Median).
INSERT FIGURE 3 HERE
Although this necessity cline represents only those expressions with positive polarity,
expressions with negative polarity realising undesirable and prohibited actions do have
a place in this continuum. Further research, however, is needed to find out how
negative directives fit in this cline; in this regard, the work of Di Eugenio (1993), and Di
Eugenio and Vander Linden (1996) offers an interesting path into the pragmatic factors
influencing the choice of what they call ‘preventative expressions’.

11
 The necessity cline can be a useful tool for translation and text-generation, but
the fact that there are several linguistic expressions in the same box shows that the
necessity cline is not self-exhaustive. If we look at the constructed examples (28) and
(29) for English, and (30) and (31) for Spanish, we can see that the necessity cline does
not distinguish between the expressions in those examples.

(28) You must connect the device to the mains.

(29) The device must be connected to the mains (by you).

(30) Conecte el aparato a la red.

(Connect [you-formal] the device to the mains.)

(31) Conectar el aparato a la red.

(To-connect the device to the mains.)

The difference between each pair of examples lies in the specification or absence of the
agent. In addition, the two English examples show differences as regards the
presentation of the information within the clause. It can then be said that two further
factors, besides action-relevance, influence the choice of directive expression. These
two additional factors are agency and theme / information.

Agency

The choices related to agency can be used (i) for providing and impersonal or detached
stance by avoiding the agent (as in (29), (31) and (32)) and (ii) for drawing attention to
the agent, as in the constructed example (33), by using ‘by’ agent.

(32) It is necessary to connect the device to the mains.

(33) The installation must be carried out by a technician.
(34) We recommend that you defrost your fridge once a week. [Hotpoint refrigerator]

Information and Theme

These refer to two informational roles of the clause (focus and theme), which arrange
the information around the two prominent positions in the clause: initial and final or
end-focus. Theme allows the instructor to draw attention - as point of departure - to
different parts of the message (for example, the agent in (28); the device in (29); the
relevance of the action in (32) and (34); or the action itself in (33)). If the agent is not
the reader (considered as default agent in instructions) but a third person, focus allows
the writer to draw attention to the agent by placing it at the end as in (33).

The networks

12
These factors can be formalised in a semantic network evolving around the systems of
ACTION-RELEVANCE, AGENCY, INFORMATION and THEME which can be a useful tool for
text generation. Figure 4 shows the network for English and Figure 5 the one for
Spanish. The curly brackets represent simultaneous choices, while the square brackets
represent exclusive choices. The linguistic realisations of each choice are represented in
italics.
INSERT FIGURE 4 HERE
INSERT FIGURE 5 HERE

It is important to note a few points about the networks:

1. The three systems - which show links with Halliday’s (1985) ideational,
interpersonal and textual functions - are simultaneous, i.e., they do not have a
hierarchical order. The choices are made at the same time, as shown by the curly
bracket notation.
2. While presented as simultaneous, there is some interdependency in the systems.
This interrelation is clearly visible in the agency system, where the choice of agent
explicitness influences the decisions of thematicity. However, the interdependency of
the systems is not shown in the networks for reasons of simplicity.
3. It is difficult to explain what makes the writer choose one option instead of the
others. For some systems, the choice is determined by the situational features (e.g.,
action-relevance). Other choices are less clear, perhaps relating to the textual context of
the instruction, or stylistic issues. A further explanation is provided by the idea that the
relationship between form and function is bi-directional (see Biber 1995: 10), which on
the other hand leads to a conundrum: the situational characteristics influence the choice
of linguistic form while the choice of linguistic features in turn helps create the
situation. Taking this into account it might be argued that my networks present
dimensions of variation rather than factors influencing choice.
4. The networks specify five different participant roles, though not all of them are
present in both networks. The roles are the following:
a) agent: the person expected to perform the action;,
b) needer: the person needing to do the action;
c) desirer: the person desiring to do the action;
d) advisee: the person advised to do the action. The data showed that this
role is relevant for the Spanish choice of directive, but not for English;
e) adviser: the person advising to do the action, i.e., the writer on behalf of
the manufacturing company.
These five different roles refer to only two participants: the writer / manufacturer, and
the user. The role of adviser is held by the writer; while the remaining roles are
generally conflated in the same person: the user, who is usually the reader. It has to be
noted that the user is not always the agent, in which case it has to be specified.

How do the networks work?

If we wanted to generate a directive such as the one in (35), several choices would have
to be made in each of the three systems. From the ACTION-RELEVANCE system, we
would choose necessity through modal, and in particular the strong modal. From the
AGENCY system, we would choose agent explicit. This latter choice would be closely

13
linked to the choice of agent-focus within the INFORMATION system, which would
render a ‘by’ agent in final position. Finally, by choosing Medium as Theme from
within the THEME system, the device would appear in initial position.

(35) The device must be installed by a technician.

If we followed the same procedure but varied one of the choices, the outcome would be
different. Thus, if we make the same choices as above for ACTION-RELEVANCE, AGENCY
and INFORMATION, but within the THEME system we choose Action as Theme instead of
Medium as Theme, the outcome would be (33), repeated here for convenience as (36):

(36) The installation must be carried out by a technician.

Finally, if instead of presenting the action as necessary we choose to present it as a

desirable action, and in particular as producer advice without specifying the adviser or
the desirer, the outcome would be the directive in (37):

(37)It is advisable that the installation be carried out by a technician.

Having seen what the main components of the networks are, the following section will
look at the networks more closely and establish the similarities and differences between
English and Spanish.

Comparing the networks

Before continuing with the comparison between Spanish and English, it must be
reminded that the networks themselves are based on the expressions found in the data
described in the section on data and method of analysis. Although I have attempted to
generalise wherever possible, other kinds of data might lead to different results and
would most likely help to expand the networks. The comparisons made in this section,
however, are strictly based on the networks for the directive choices of presented earlier
in Figure 4 and Figure 5.
The networks for English and Spanish are completely similar in the systems of
INFORMATION and THEME. Although the other two systems (ACTION-RELEVANCE and
AGENCY) also show great similarities, the differences are worth noting and will be
summarised next.
The system of ACTION-RELEVANCE presents the following differences between
both languages:
• The necessity marking is more complex in English than it is in Spanish. In
particular, English presents a choice between two modals, while the Spanish data only
has one modal available for expressing directives in instructions.
• Within the desirability marking, the producer’s advice is more complicated for
Spanish than it is for English, allowing both advisor and advisee to be potentially
included.13
When comparing the networks, however, it can be seen that the main difference
between Spanish and English concerns the AGENCY system, which is much more
complex for Spanish. The Spanish network presents options for agent-markedness,

14
agent-inclusion and distance, within the agent-explicit option, while English simply
distinguishes between agent-implicit and agent-explicit, without any further options in
the latter. The most interesting conclusions that can be drawn from the AGENCY system
are specified in the following. In some situations the instructor will prefer not to
address her audience directly, thus leaving the actor unspecified; this is what I will call
‘agent defocusing’, because it draws the attention away from the agent. The following
are choices of agent defocusing available to both Spanish and English:
• Use of passive voice in modalised directives. This has an effect on information
structure as well.
• Impersonal expressions of appeal to the reader (e.g., it is advisable to).
There are also some ways of agent defocusing exclusive to Spanish:
• The use of the infinitive as a directive.
• Impersonal passive imperative with se.
• Use of the plural instead of the singular (e.g., soliciten [‘you’ formal-plural] instead
of solicite [‘you’ formal-singular]).14
• Finally, a further Spanish mechanism of indirectly addressing the reader is the use of
the 1st person plural instead of the 2nd person singular. This is a form of indirect
address because the writer involves herself in the action performance.
To finish, one of the most interesting issues is the effect that agency and
explicitness of advice can have on the relationship that the writer / manufacturer
establishes with the reader: by specifying the advisee or the agent the writer provides a
friendly relationship. Agentless expressions in contrast provide a telegraphic style and
tend to be interpreted as more distant in tenor.

Conclusions

This paper addressed one of the main linguistic components of the instructional genre:
directive expressions. The imperative clause is thought to be the syntactic category that
typically realises a directive. However, it was shown that a wide range of other
expressions are also used. The aim of this paper, then, was to explain the availability of
so many different directive expressions in instructions.
Taking into account the definitions and classifications of directives in the
literature, the second section offered a description of what is meant by ‘directives’ in
instructions: expressions aimed at getting the reader to do (or not to do) the actions in
the task-plan. It was also discussed that directives in this particular genre are hearer-
interested in the sense that the actions directed are of benefit to the addressee. Only
direct directives were considered in this study, though the existence of indirect
directives (e.g., hints, question directives, etc.) in the data was acknowledged.
The paper drew upon Biber’s (1995) suggestion that situational features are
related to linguistic forms through functional and conventional associations. Section 5,
therefore, showed that situational features explain the appropriateness for instructions of
directive forms, such as the imperative and modalised expressions with must, which are
considered impolite in other contexts. It was argued that syntactic expressions do not
have a fixed politeness level, but their degree of politeness rather depends on contextual
features such as mode of communication, authority relations and task-relevance in the
following ways:

15
 The same directive expression is interpreted in different ways in speech and
writing. It was argued that the imperative, which is usually perceived as rude in speech,
is considered appropriate in instructions because the action it expresses is of benefit to
the addressee.
It was argued that the power relations involved in written instructions do not
have to do with rank or social status, but rather with the authority coming from the
instructor’s superior knowledge. This is what enables the writer to use straight
directives and modalised directives with must, which would be considered impolite in
other situations.
Instructions are task-relevant and therefore allow for the use of direct directives,
avoiding the use of embroidered directives that would draw the reader’s attention away
from the task-plan.
It was argued that the wide range of expressions is due to the fact that each
expression conveys different pragmatic and semantic meanings. The factors influencing
the choice of directive expression in English and Spanish were presented graphically
through two networks (Figures 4 and 5), which reveal the close relationship between
context and linguistic form. The networks identify four simultaneous and intertwined
systems of options (ACTION-RELEVANCE, AGENCY, INFORMATION and THEME) which
influence the choice of expression. The networks can be a useful tool for automatic
generation of instructional texts.
It is obvious that not all the actions in the task-plan are equally important: there
are actions which are strictly necessary; other actions, though not necessary, are
particularly beneficial to the addressee, and therefore they are desirable. It can be said
that the actions in the task-plan present a hierarchy depending on their role in the goal of
the task-plan itself (i.e., on using the product efficiently). The actual form of directive
used conveys information about the relevance in the task-plan of the action it
introduces. This is marked through the system of ACTION-RELEVANCE. Taking into
account that unmarked actions such as the ones realised by the imperative are
obligatory, it was argued that directive expressions can be explicitly marked for one of
the following categories: necessity, desirability, undesirability, and prohibition.
The system of AGENCY involves the explicit mentioning of the agent of the
actions. The agent in instructions is by default the reader and, since it can be recovered
from the context, it is not always necessary to specify the agent. The ways of agent
defocusing and the effect they have in the relationship established between the writer
and the reader were also shown in the section comparing the networks.
Closely linked to agency are the systems of THEME and INFORMATION. These
systems reflect the way the writer organises her message taking into account the
prominent locations in the clause. Thus, the writer can draw the addressee’s attention to
either the action itself or the Medium (most generally the device or one of its parts) by
placing them in thematic position. Focus, on the other hand, allows the writer to draw
attention to the agent, when this is not the reader himself but a third person, by locating
it in the end-focus position.

16
References

Agre, P. E. and Batali, J. 1991. “Collaborative Instructional Use in Furniture

Assembly”. In N. Gilbert (ed.), Proceedings of the AAAI Workshop on Complex
Systems, Ethnomethodology and Interaction Analysis. Boston, 1991, 1-16.
Austin, J. L. 1962. How to Do Things with Words. Oxford: Clarendon Press (2nd
edition).
Biber, D. 1995. Dimensions of Register Variation. A cross-linguistic comparison.
Cambridge: CUP.
Brown, P. and Levinson, S. 1987. Politeness. Some universals in language usage.
Cambridge: CUP.
Butler, C. S. 1988. “Politeness and the semantics of modalised directives in English”. In
J. D. Benson, M. J. Cummings, and W. S. Greaves (eds), Linguistics in a Systemic
Perspective. Amsterdam: John Benjamins, 119-153.
Butt, J. and Benjamin, C. 1988. A New Reference Grammar of Modern Spanish.
London: Edward Arnold (2nd edition).
Crystal, D. 1992. An Encyclopaedic Dictionary of Language and Languages. Oxford:
Blackwell.
Delin, J., Scott, D. and Hartley, A. 1993. “Knowledge, Intention, Rhetoric: Levels of
Variation in Multilingual Instructions”. In O. Rambow (ed.), Intentionality and
Structure in Discourse Relations. Proceedings of a Workshop sponsored by the Special
Interest Group on Generation of the Association for Computational Linguistics. 21 June
1993. Ohio State University, Columbus, Ohio, 7-10.
Di Eugenio, B. 1993. “Speakers’ Intentions and Beliefs in Negative Imperatives”. In O.
Rambow (ed.), Intentionality and Structure in Discourse Relations. Proceedings of a
Workshop Sponsored by the Special Interest Group on Generation of the Association
for Computational Linguistics. 21 June 1993. Ohio, State University, Columbus, Ohio,
11-14.
Di Eugenio, B. and Vander Linden, K. (1996). “A Corpus Study of Negative
Imperatives in Natural Language Instructions”. In Proceedings of the 16th International
Conference on Computational Linguistics, COLING96. Copenhagen, Denmark, 346-
351.
Dixon, P. 1982. “Plans and Written Directions for Complex Tasks”. Journal of Verbal
Learning and Verbal Behavior 21: 70-84.
Dixon, P. 1987a. “Actions and Procedural Directions”. In R. Tomlin (ed.), Coherence
and Grounding in Discourse. Amsterdam: John Benjamins, 69-89.
Dixon, P. 1987b. “The Structure of Mental Plans for Following Directions”. Journal of
Experimental Psychology: Learning, Memory and Cognition 13 (1): 18-26.
Dixon, P., Faries, J. and Gabrys, G. 1988. “The Role of Explicit Action Statements in
Understanding and Using Written Directions”. Journal of Memory and Language 27:
649-667.
Downing, A. and Locke, P. 1992. A University Course in English Grammar. Hemel
Hempstead: Prentice Hall.
Ervin-Tripp, S. 1976. “Is Sybil there? the structure of some American English
directives”. Language in Society 5 (1): 25-66.
Fairclough, N. (ed.). 1992. Critical Language Awareness. London: Longman.
Halliday, M. A. K. 1978. Language as Social Semiotic. The social interpreation of
language and meaning. London: Edward Arnold.

17
Halliday, M. A. K. 1985. An Introduction to Functional Grammar. London: Edward
Arnold (2nd edition).
Hartmann, R. R. K. 1980. Contrastive Textology: Comparative Discourse Analysis in
Applied Linguistics. Heidelberg: Julius Groos Verlag.
Hartmann, R. R. K. 1996. “Contrastive Textology and Corpus Linguistics: On the Value
of Parallel Texts”. Language Sciences 18(3-4): 947-957.
Hasan, R. 1978. “Text in the Systemic-Functional Model”. In W. Dressler (ed.). Current
Trends in Text Linguistics. Berlin: de Gruyter, 228-246.
Huddleston, R.D. 1984. Introduction to the Grammar of English. Cambridge: CUP.
Paris, C. and Scott, D. 1994. “Stylistic Variation in Multilingual Instructions”. In
Proceedings of the 7th International Workshop on Natural Language Generation.
Kennebunkport, Maine, 1994, 45-52.
Real Academia Española. 1973. Esbozo de una nueva gramática de la lengua española.
Madrid: Espasa-Calpe.
Ross, A. 1968. Directives and Norms. London: Routledge & Kegan Paul.
Sacerdoti, E.D. 1977. A Structure for Plans and Behavior. New York: Elsevier North-
Holland.
Sager, J. C., Dungworth, D. and McDonald, P. F. 1980. English Special Languages.
Principles and practice in science and technology. Wiesbaden: Brandstetter Verlag.
Searle, J. R. 1969. Speech Acts: An essay in the Philosophy of Language. Cambridge:
CUP.
Searle, J. R. 1976. “The classification of illocutionary acts”. Language in Society 5 (1):
1-24.
Searle, J. R. 1979. Expression and Meaning. Cambridge: CUP.

1
The term register in Biber (1995) corresponds to what is commonly described as
genre.
2
A similar claim has been made by Halliday (1978: 117) and Fairclough (1992: 8),
among others.
3
For the complete classification and explanation of these categories see Ross (1968:
34-60).
4
Only one example of imbedded imperatives such as “We would ask you to ...” in
each language was found in the data.
5
Note that Ervin-Tripp’s imperative as a kind of directive includes modalised
statements as well as imperative forms. According to her, the following two examples
would be classified as imperatives:
The bowl should be washed and dried after use and before storage. [Boots kitchen
scale]
Siempre debe cubrir los alimentos a freír. [Jata deep fryer]
(You must always cover the food for frying.)
To avoid confusion, from now on every time the term “imperative” is used it will refer
to the linguistic realisation rather than the directive type, unless otherwise stated. I,
therefore, do not use imperative in Ervin-Tripp’s sense, but rather distinguish between
imperative and modalised realisations.
6
It will be normal procedure throughout the paper to italicise the original examples
(i.e., the examples found in the data), as well as to provide their source in square brackets.
Constructed examples, however, will be presented in courier font.

18
7
Polarity is an in-built feature of the grammar and it did not seem relevant to
indicate it in the networks. However, a specific semantic network for negative directives
might be relevant and will require further research.
8
It will be normal procedure throughout the paper to provide a gloss, rather than a
translation, together with the Spanish examples.
9
Spanish prescriptive grammars such as Real Academia Española (1973) have
always rejected the use of the infinitive as a directive. However, the infinitive use as
directive in this kind of data is totally grammatical and its use is spreading, as pointed out
by Butt and Benjamin (1988: 278).
10
Translation: “Open the grill and place the food for cooking.”
11
It will be normal procedure in this paper to use the feminine pronoun (she) to refer
to the writer, while the masculine pronoun (he) will refer to the reader.
12
See note 10.
13
If the networks were extended to cater for directives in other kind of data, the
producer’s advice system in English would surely present more similarities with Spanish,
since it would be possible to find “we advise you to ...”
14
In English, the third-person plural can also be used impersonally, as in the two
examples below taken from Butt and Benjamin (1988: 374). English, however, cannot use
this form with the imperative.
They say exercise is good for the heart.
It seems that they speak more slowly in the USA than in England.

19