Expressing Procedural Relationships in 
Multilingual Instructions* 
Judy Delin 
Dep~rtment of English Studies 
University of Stirling 
Stifling FK9 4LA, UK 
email: J.L.Delin'~stirling.ac.uk 
Anthony Hartley C4cile Paris Donia Scott Keith Vander Linden ; 
Information Technology Research Institute 
University of Brighton 
Lewes Road 
Brighton BN2 4AT, UK 
email: { Tony.Hartley, Cecile.Paris,Donim Scott,I~eith.Vander-Linden } @it ri.bton.ac.uk 
Abstract 
In this paper we discuss a study of the expression of 
procedural relations in multilingual user instructions~ in 
particular the relations of Gene,ution and Enablement. 
These procedural relations are defined in terms of a plan 
representation model, and applied in a corpus study of 
English, French, and Portuguese instructions. The re- 
sults of our analysis indicate specific guidelines for the 
tactical realisation of expressions of these relations in 
multilingual instructional text. 
Introduction 
In user instructions, it is common to find expressions 
like: 
(1) Pull down and remove the white plastic tray that 
holds the video cable and unpack the cable.(Apple) 
Here we have what appears to be a simple expression 
of sequential actions. The reader is expected to pull 
down the white plastic tray, remove that tray: and then 
unpack the cable. Current studies of instructional text 
(e.g.. RSsner and Stede, 1992; Vander Linden, 1993) typ- 
ically represent the relationship found here in a simple 
multi-dependent structure such as that provided by the 
Sequence schenla in Mann and Thompson:s Rhetorical 
Structure Theory (RST) (1988) (see figure 1A). This 
rhetorical structure, which represents three sequential 
actidns, directly matches the grammatical forms used in 
the actual text. Indeed: this sequential execution is pre- 
cisely what the instruction writer desires the reader to 
"This work is supported by the Commission of the Euro- 
pean Union Grant LRE-62009, the Engineering and Physical 
Sciences Research Council Grant 319221, and the BC/DAAD 
ARC Project 293. 
tAuthors are presented in alphabetical or(ler. 
do in this context. The complication comes when one 
finds, sometimes even in the same instruction mamlal, 
an alternate form of expression for the very same user 
action; such as the one shown in the following: 
(2) Pull down and remove to unpack the video cable. 
(Apple) 
R ST analysts would most likely represent this with a 
purpose relation, as in figure lB. Clearly, the represen- 
tation formalism fails to capture the common semantics 
of these two examples. 
This problem may be only rarely encountered in a 
monolingual context, but is exceedingly common in a 
multilingual enviromnent (Delin et al., 1993). Consider 
the following example, taken from a trilingual, trans- 
lated instruction nlanual for a step-aerobics machine: 
(3) The stepping load can be altered by loosening the 
locking lever and changing the position of the cylinder 
foot. (Liftmaster) 
(4) Pour modifier la charge d;appui, desserrer 
To modify 
les leviers puis 
the levers then 
v~rins.(Liftmaster) 
cylinder foot 
(5) Nach Lockern 
After loosening 
the load stepping loosen 
d6placer le pied des 
change the foot of the 
der Klemmhebel kann 
of the levers can 
durch Verschieben des ZylinderfuBes die 
by pvzhing of the cylinder foot the 
Tretbelastung ver~ndert werden.(Liftmaster) 
load changed be. 
Here we find the same information being conveyed 
through expressions that exhibit rather different RST 
61 
7th International Generation Workshop • Kennebunkport, Maine • June 21-24, 1994 
A. Example 1 
Pull-Down Remove Unpack 
B. Example 2 
Seque~ Unpack 
Pull-Down Remove 
Figure 1: Contrasting Discourse Structures Representations 
structures (see Figure 2). The English version is per- 
haps best characterised as a Means relation, whereas 
the French is a Purpose relation and the German a Cir- 
cumstance. 
The problem examples such as these raise for genera- 
tion is twofold: (a) how do we endow our systems with 
the necessary flexibility to generate alternative expres- 
sions of the same basic information and (b) given this 
flexibility., how do we determine which of the available 
expressions are appropriate? The root of these problems 
is that, to date, research in the generation of instructions 
has failed to take account of the precise procedural re- 
lations that underlie the texts. In this paper, we make 
an excursion into the philosophical literature on Action 
Theory and the AI literature on plan recognition to find 
solidly defined procedural relations which can capture 
the underlying semantics of actions and their relation- 
ships in a way that will prove useful for our generation 
task: building multilingual instructional text generation 
systems. We present the results of part of an ongoing 
investigation of the utility of the Gene~ution and En- 
ablement relations in contributing to our stated goal. 
Procedural Relations 
The philosopher Alvin Goldman identified two basic 
procedural relations; Generation and Enablement~ that 
apply between actions in a task (GoMman: 1970). These 
have been formalised within the Planning community 
and shown to be useful constructs for plan recognition 
(Pollack, 1986; Balkanski: 1993). In this section we re- 
view the definitions of Genelution and Enablement and 
describe how we have applied them in our analysis of 
a corpus of instructional texts in English, French and 
Portuguese. 
The Generation Relation 
A simple test of Generation holding between action pairs 
is whether it can be said that by performing one of the 
actions (a); the other (/3) will automatically occur (Pol- 
lack: 1986); if this is the case then it can be said that 
(a) gene~utes (/3). According to Goldman, Generation 
is a relation applying between two aet-token.~ in certain 
well-defined situations. For purposes of producing in- 
structions, however, we find it necessary to view Gener- 
ation as hohting between act-types: since we are dealing 
with actions that have not yet taken place; we are able 
to do this without distorting the general theory that 
Goldman presents (cf. Pollack. 1986). Here are some 
canonical examples of this relation taken from Goldman: 
(6~a) John flips the switch. (cQ 
(b) John turns on the light. (/~) 
where flipping the switdl generates turning on the light 
and 
(7~a) John moves his queen to king-knight-seven. (cr) 
(b) John checkmates his opponent. (/3) 
where moving the queen to king-knight-seven generates 
checkmating the opponent. 
Following Goldman, Generation holds in cases where: 
• the actions concerned are performed by the same 
agent (e.g., John). For reasons that will become 
dearer later on, we have found it necessary to qual- 
if)' and extend this part of the definition to actions 
performed or perceived to be pelJolTned by the same 
human agent. 
• the relation is asymmetric: if the agent does /3 by 
doing a' then he does not do a' by doing/3 (e.g., the 
switch cannot be flipped by turning on the light). 
• neither action is subsequent to the other (e.g.; the 
light does not come on after the switch is flipped, or 
~rice versa) l . A useful linguistic test for this is the 
invalidity of the expression '% and then i3". 
• the actions are not co-temporal. Simply stated, if 
a and 3 are co-temporal, then one could validly say 
"c~" while also /3" and vice versa: for example, "John 
strummed the guitar while also wriggling his toes". 
These actions are co-temporal, but not related by 
Generation. One could not validly say that John 
flipped on the switch while also turning on the light, 
or moved his queen to king-knight-seven while also 
checkmating his opponent. 
• if the agent had not done a, then he would not have 
done/3; 
• there is a (set of) condition(s) "i (say, George making 
sure that the light bulb was screwed in) such that 
According to Goldman, this holds even if the light does 
not come on for a few seconds at'ter the switch was flipped. 
62 
7th International Generation Workshop • Kennebunkport, Maine * June 21-24, 1994 
A. Example 3 (English) 
Alter ~quence 
Loosen Change 
B. Example 4 (French) 
Loosen Change (Desserrer) (Deplaeer) 
C. Example 5 (German) 
Loosen \[ (Lockern) 
Clmnge Alter (Versclueben) (Verandert) 
Figure 2: Contrasting Multilingual Discourse Structures Representations 
- the conjunction of a and 7 entails/3; 
- if 3" had not obtained: the performance of a' would 
not have led to/3. 
A GeneTution relation between action pairs can re- 
ceive a variety of linguistic realisations. Consider the 
following semantically equivalent instructions: 
(8) Checkmate your opponent by moving your queen to 
king-knight seven. 
(9) Move your queen to king-knight seven to check, rate 
your opponezkt. 
(10) I~ you want to checkmate your opponent, move your 
queen to king-knight seven. 
Two points are worthy of note here. The first is that 
although causation plays a key r61e in GeneTution, /3 is 
not in a strict sense caused by a'; rather,/3 is the action 
of (to use Goldman:s term) 'bringing about' the state 
that is the direct result of performing cr under specific 
conditions (7)- Referring back to example 6: John's flip- 
ping of the switch causes the event of the light coming 
on; John's 'bringing about' of this event can be captured 
by the expression 'turning on the light:. 
The second point to be noted is that in the domain of 
instructions for mechanical and electronic objects there 
is also a hidden agent the machine itself, as an agent 
acting on behalf of the user. This leads to the possibility 
of equivalence of expressions such as ':I deleted the file" 
and ':My coml)uterdeleted the file". 'We consider these 
to be equally valid expressions of the generated part of a 
Generation relation. For example, we also count instruc- 
tions such as the following (for using a cooking utensil) 
(11) Imprimez un mouvement de rotation "~ 
Impart a movement of ~vtation to 
la garniture: elle se d~.nlonte.(Darty) 
the plate, it itself removes. 
as instances of Generation. Expressions of generated 
actions with non-human agents are relatively uncommon 
in English instructions but (as will be discussed later) 
this is not true of all languages. 
The Enablement Relation 
Instructions also inchtde expressions such as the folh)w- 
ing: 
(12) You must restart your Macintosh to use your new 
software.(Mierosoft) 
(13) Before using your data cartridge: allow it to acch:- 
mate to your operating environment for eight hours 
or fo7" the amount of time it has been exposed to dis- 
similar conditions whichever is less.(3M) 
(14) 2 Double-click tl~e Installer icon with the mouse 
button. 
The installer displays an introductory screen. 
3 Click the OK button.(Microsoft) 
(15) In order to operate remote control: you m,u, st first 
have noted or" memorised your code.(BeU) 
These are clearly not instances of Generation, since 
the execution of any member of the action pair does 
not 'automatically ~ result in the execution of the other. 
In each case: some intermediating action(s) is required. 
Following Goldman (1970), this type of relation between 
action pairs is most commonly referred to as Enablement 
(Pollack, 1986; Balkanski: 1993; Di Eugenio, 1992). 
Goldman does not provide a formal definition of this 
relation, but Pollack gives a useful general definition: " 
... when a enables/3, then the agent needs to do some- 
thing more than o to guarantee that/3 will be done." 
(Pollack, 1986: page 40). A more detailed definition is 
provided by Balkansld: ': ... Action Ai enables action 
A3 if and only if there is a set of conditions: C, such that 
one of the conditions in C: Ci: holds as a result of the 
performance of Ai: and either there is a third action A~ 
such that A2 conditionally generates A3 under C; or C 
is the executability condition of A3." (Balkansld: 1993, 
page 41). 
Enablement applies to all four examples above because 
in them: one action brings about conditions that are nec- 
essary: but not necessarily sufficient for the subsequent 
performance of the other. In eexample 12: the reader 
must restart the Macintosh in order to be in a position 
to be able to use the new software. There are in fact 
other actions necessary for the use of the software, but 
they are not mentioned. Acclimating the cartridge in 
example 13 is not sufficient to make it in use. Similarl)~ 
in example 14, clicking on the Installer icon only brings 
about a set of conditions (in this case, the appropriate 
menu), for the user to click the OK button. In example 
63 
7th International Generation Workshop • Kennebunkport, Maine • June 21-24, 1994 
15: the reader nmst know the code in order to be able to 
operate the remote control: although knowing the code 
is not enough. 
Note that in some cases of Enablern.ent (e.g.: in 12 and 
13)~ the intermediate action can in fact be performed in 
the absence of the enabling condition; the result will 
simply be that the enabled action will not occur. In 
others however (e.g.: examples 14 and 15), the interme- 
diate action is impossible to perform when the enabling 
condition does not hold. 
An important property of Enablement which distin- 
guishes it from Generation is that the enabled and en- 
abling actions can be carried out by different human 
agents. Goldman alludes to this in his discussion of ex- 
ample 6: where he introduces a necessary condition for 
the Generation of John turning on the light: George's 
screwing the bulb into the socket. 
As in the case of some of the Generation examples 
we saw before, actions in Enablement relations can be 
expressed as an activity carried out by a human agent or 
by. say: a machine or in terms of the associated state of 
the agent. For example, example 15 could equally have 
been expressed as the following: 
(16) In order teJ operate remote control: you must know 
your code. 
(17) For the remote control to operate, you must first 
have noted or memorised your code. 
A Formalisation 
From our discussion so far, it is clear that the rela- 
tions of Gene,ution and Enablement should be identi- 
fiable aspects of plans for perfornfing procedures. We 
employ a plan formalism that is a simple extension 
of STRIPS-styled operators developed by Fikes (1971) 
and expanded in the NOAH system (Sacerdoti: 1977). 
In our representation, plans comprise: 
goal(s): an action (or set of actions) which motivate(s) 
the use of the plan; 
constraints: states which nmst hoht before a plan can 
be employed. Constraints cannot be achieved through 
planning; 
side-effects: states which arise as unplanned effects of 
carrying out a plan; 
a body: an action or action complex which executes 
the plan; if these are not l/rimitive: they can them- 
selves be achieved through another plan; 
preconditions: an action or action complex which, 
when carried out, leads to conditions necessary for the 
successful execution of the plan; i.e: the body will be 
executable but its execution will not generate the goal 
(even if the constraints hold) unless the precondition 
is realisable. Preconditions can be planned for} 
-"This type of preconditions relates to Pollack's (1986) 
g~neration-enablin.q ~:ondition. 
Plans must minimally have a body and goal. Actions: 
in turn, comprise: 
constraints: these have the same properties as the con- 
straints on plans; 
side-effects: these also have the same properties as the 
side-effects on plans; 
effects: states which arise from the bringing-about of 
the action; 
preconditions: like the preconditions of plans; these 
are either an action or an action complex and can be 
planned for. Unlike the preconditions of plans, how- 
ever, their effects must hold before the action can take 
place (i.e., the action cannot result if the precondition 
does not hold). 3 
Actions must minimally have an effect. Finally: states 
can have evidences. These are phenomena which signify 
that the state holds. 
As in all STRIPS-based planning formalisms, plans 
in our representation can have associated sub-plans. In 
our case: sub-plans arise through the body of a plan: the 
preconditions of a plan: or through the preconditions of 
an action. 
Having set out this scheme, we can now proceed to 
identify instances of Generation and Enablement within 
our representation for instructions: 
a generates ~ iff a' is the body of a plan e whose 
goal is B. 
a enables/) if a is a precondition of a plan e and 
is the goal of plan e: or if/3 is the body of e and a' 
is a precondition of/3. 
Figure 3 shows graphically the relationship between 
the notion of a plan and the Generation and Enablement 
relations. 
Generation and Enablement in 
Instructions 
\'Ve have shown through examples that Generation and 
Enablement can be expressed in a variety of ways. In in- 
struction generation, it is clearly important to signal the 
correct relationship between actions, since failure to do 
so could lead to situations where the user diligently car- 
ries out the action without achieving the desired effect. 
For example, if the following Enablement relation 
(18) To clean the case: disconnect the power 
plug.(Microsoft) 
is misperceived as Generation (as it might be were it 
to be expressed as "Clean the case by disconnecting the 
power plug"): the user would believe that by disconnect- 
ing the plug she would be guaranteed to achieve her goal 
of cleaning the case; this would only result in frustra- 
tion, bewilderment and a still dirty case. We are thus 
:~This Lotion is closely related to Balkanski's (1993) deft- 
nition of ezecutabiIity condition. 
64 
7th International Generation Workshop • Kennebunkport, Maine ° June 21-24, 1994 
,- I 
," / \ I 
! ....----I ,,1,,o L .... . : , 
, , ; 
Relations: 
Generation 
1111111111011111111 
Enablement 
/" Precondition | / 
Go-To(Switch) \] 
Entities: 1 Actions \] I 
States 
Plans ~ 
Figure 3: Generation and Enablement in the Plan for Turning on the Light 
concerned to generate only the appropriate expressions 
for Generation and Enablement. 
Our approach to this problem is to look to the data: 
we are analysing a corpus of commercially-produced in- 
structions for manual tasks to extract the mappings be- 
tween Generation and En.ablement and linguistic form. 
The results of our study inform our design of a system 
for generating draft instructions in multiple languages 
front a common; language-independent: representation. 
~'Ve describe some of our results in this section, based on 
our analyses of English: French and Portuguese instruc- 
tions. 
The data we report are from original texts written in 
the particular language being addressed; they are not 
the product of translation. Our method of analysis is to 
identify instances of Gene~utiou and Enablement from 
the corpus and to register its grammatical form. We 
report here only on data for which we all agree on the 
chosen procedural relationship. 
In analysing the data, our main concern has been to 
determine: 
1. What are the commonly used mappings front Gener- 
ation and Enablement onto language? 
2. Which forms are ambiguous between Generation and 
Enablement and how can we exploit this information 
for generation? 
3. What are the factors that influence the 'appropriate- 
ness: of a selected form? 
The English Results 
The English corpus is made up of 451 clauses (approx- 
imately 3500 words) taken from 9 instruction manuals 
intended for non-expert readers. A detailed accounting 
of the grammatical forms used to express Generation 
relations is shown in table 1. Similar statistics for En- 
ablement relations are shown in table 2. 
The general results of these mappings are shown 
graphically in figure 4, and as can be seen: there are 
grammatical forms which tend to be exclusively used for 
either Genelution or Enablement (imperative + by, sequence, 
and before conditions) and other forms that 
are common to both relations (to infinitives and :for 
phrases). 
The most common forms shown in tables 1 and '2 ap- 
pear to be clear patterns of use which can be exploited 
in text generation, rather than just random selections 
or artifacts of this particular corpus. 
Expressing Generation in English In our corpus: 
GeneTution relations tend to be expressed with either the 
imperative + by: the imperative + to infinitive, 
or one of a set of :for phrases. These forms, which to- 
gether constitute 80% of the Gene~zttion relations iden- 
tiffed, are exemplified here in examples from manuals 
for a Macintosh colour monitor and for a Bell telephone 
answering machine: 
(19) If your computer system does cause interference to 
radio or television reception: try to correct the inter- 
ference by using one or more of the following men- 
65 
7th International Generation Workshop • Kennebunkport, Maine • June 21-24, 1994 
Functions 
Forms 
Figure 4: The English flmction to form mapping of Generation and Enablemeat Relations 
Grammatical Form Count Percentage 
To Infinitive 30 46.1% 
For Phrases 14 21.5% 
Imperative + by 8. 12.3% 
Clause Sequences 3 4.6(~, 
7 other forms 10 15.4% 
Total: 65 
Table 1: The frequency of English grammatical forms 
for Generation Relations 
Grammatical Form Count Percentage 
Sentence. Sequence 22 47.8% 
To Infinitive 10 21.7(/o 
For Phrases 5 10.9(/0 
Before Conditions 4 8.7% 
Imperative + by 1 2.2(/t, 
4 other forms 4 8.7(/(, 
Total: 46 
Table 2: The frequency of English grammatical forms 
for Enablement Relations 
sures:(Apple) 
(20) To interrupt playback: rotate the VOLUME knob 
towards the rear, past the 'click:.(Bell) 
(21) See your authorized Apple dealer for information 
on screen save programs.(Apple) 
In example 19, the action of using "one or more of the 
following measures" generates the action of correcting 
the interference. They are performed by the same hu- 
man agent, asymmetric, one does not occur before the 
other, and there exist the specified conditions -f; the 
same pattern is repeated in examples 20 and 21. 
The first question is what to make of the gaps? Why 
are Gene,ution relationships so infrequently expressed 
using sequences of clauses, and never expressed using 
before clauses? This clearly follows from the definition 
of Generation that neither action can be subsequent to 
the other, a relationship that is strongly implied by se- 
quences. There are cases where sentence sequences 
are used to express Generation, but these are infrequent 
and occur in very specific syntactic situations (Vander 
Linden, 1993). In general; grammatical forms which 
typically function as expressions of sequence should be 
avoided when expressing Generation relationships be- 
cause they tend to mislead the reader by suggesting the 
performance of two distinct actions. 
Concerning the question of identifying the contexts in 
which the imperative + by, the to infinitive, and 
the for phrase arise, it has been observed that syntac- 
tic constraints and lexical availability appear to moti- 
vate the use of the imperative + by form and the for 
phrasal forms (Vander Linden, 1993). 
Expressing Enablement in English Enablement 
relations tend, in 89% of the cases, toward before con- 
ditions: one of a set of sequence forms; imperative + to 
infinitive and other for phrasal types. Examples of 
these forms are given here: 
(22) Quit all applications before beginning the following 
procedure.(Microsoft) 
(23) Switch offat the mains, disconnect the power adap- 
tor and remove the cassette. Wipe the tape heads and 
pinch roller gently with a cotton-wool bud dipped in 
tape head cleaning fluid or methylated spirits.(Bell) 
(24) Return to seat to place calls.(Airfbue) 
(25) PLEASE SAVE THESE INSTRUCTIONS FOR 
FUTURE USE.(Playschool) 
In example 22, quitting all applications is an action in- 
tended to bring about a set of conditions compatible 
with "beginning the foUowing procedure": as are the ac- 
tions of removing the cassette, returning to the seat, and 
saving these instructions in examples 23, 24, and 25 re- 
spectively. None of them are sufficient to achieve their 
intended goal. 
The gap in the expressions of Euablement is that they 
are seldom expressed using the imperative + by form. 
As can be seen in tables 1 and 2; there is only one En- 
ablement expressed as an imperative + by as opposed 
to eight Gene~ution relations expressed that way. It 
seems inappropriate, for instance, to express the actions 
in example 24 as "Place calls by returning to your seat" 
because it tends to imply that returning to your seat is 
all you have to do to achieve the goal. We, thus, would 
restrict the use of the imperative + by to the expression 
of Genelution relations only. 
The French Results 
The French corpus comprises 14 sets of instructions 
66 
7th International Generation Workshop • Kennebunkport, Maine • June 21-24, 1994 
Func~on 
Forms 
Figure 5: The French function to form mapping of Generation and Enablement Relations 
Grammatical Form Count Percentage 
En-present-participle 12 33.3% 
Pour-infinitive 8 22.2% 
Apposed clause 5 13.9% 
Pour-noun 4 11.1% 
Afin-de-infinitive 4 11.1% 
3 other forms 3 8.3% 
Total: 36 
Table 3: The frequency of French grammatical forms for 
Generation Relations 
Grammatical FFoTw~ Count Percentage 
Et-sequence 13 40.6% 
Pour-Infinitive 6 18.8% 
Avant-conditions 6 18.8% 
Apr~s-conditions 4 12.5% 
2 other forms 3 9.4% 
Total: 32 
Table 4: The frequency of French 
Enablement Relations 
grammatical forms for 
h,r French-produced domestic appliances and for home 
improvement. Thus, the texts are intended for non- 
specialists. Most have a word-count of between 200 and 
400 words. We set a target of collecting some 30 ex- 
amples of each relation, expressed in single sentences 
rather than across sentence boundaries. The examples 
were vetted by a native speaker. 
The statistical analysis of surface forms is given in 
tables 3 and 4. The graphical representation of these 
data: shown in figure 5, indicates only one form to be 
common to both relations. 
Expressing Generation in French French has a re- 
alisation of this relation that is distinct from English, as 
we showed in example 11: and can see again here in an 
instruction from a Philips telephone manuah 
(26) Appuyer 2 fois sur la touche MEMO: 
PTvss 2 time.~ on the key Mt~MO, 
le nun~ro se conxpose alors 
the nv.mber itself dials then 
au tomatiquement.(Philips) 
automatically. 
Here the information is expressed through clauses which 
are in apposition. The same information could have 
been expressed as: 
(27) Appuyer ... sur la touche pour 
Press ... on the key for 
composer ... 
to dial ... 
It is examples such as these that led us to modify 
Goldman's definition of GeneTution. In French it is com- 
mon to achieve agent deletion by using a reflexive verb 
whose syntactic subject is the semantic theme of the cor- 
responding fifll transitive verb. Although there is no ex- 
plicit agent of the action, agency is readily inferred and 
the spirit of Goldman;s definition is preserved. More- 
over: the word "automatically" is in a sense redundant, 
serving to emphasize the fact that the result is achieved 
with no filrther action on the part of the agent. 
Agent deletion may also be achieved by de-verbal 
nominalization: as in the following examples from man- 
uals for a cooking utensil and for do-it-yourself home 
improvement: 
(28) La pouss6e du ressort interne d6gage 
The push of the spring inteTnal releases 
le poussoir.(Darty) 
the push button. 
The most frequent realization is by the en present 
partic±ple form, for example in: 
(29) Soulever manuellement le ttotteur au 
Rai.se manually the .float to the 
maxinmm en tirant sur le 
maximum by pulling on the 
bras.(Castorama 50.11) 
aTT/L 
Although it is tempting to view this form as the direct 
equivalent of ':by + ing', it is also available for relating 
distinct co-temporal actions: with no implication that 
one is achieved by execution of the other: as in the fol- 
lowing instruction for a jar opener: 
67 
7th International Generation Workshop • Kennebunkport, Maine • June 21-24, 1994 
(30) Serrer la poign~e de blocage d:une 
Tighten the blocking-handle with one 
main en maintenant le bocal de 
hand while holding the jar with 
l'autre main.(Tefal) 
the other" hand. 
The next example illustrates a further common real- 
ization using "pour" which reinforces the observa- 
tion that generated actions and evidences appearing in 
the same sentence are expressed differently: 
(31) S~lectionner une touche du clavier (de 
Select a key of the pad (f~vm 
1 ~t 9) pour rendre la ligne (la tonallt~ 
1 to 9) to return the line (the tone 
revient).(Philips) 
returns). 
It is worth noting that, although "pour" and "afin de" 
are generally considered interchangeable in this syntac- 
tic construction, the latter is: in our corpus; used exclu- 
sively to express Generution while "pour" realizes both 
procedural relations. 
Expressing Enablement in French It is in the ex- 
pression pour in:firtitive that Generation and Enable- 
ment overlap. However, the construction so strongly sig- 
nals Generution for French readers that they may have 
difficulty construing an Enablement relation: 
(32) D6.visser le flotteur de son bras pour 
Unscrew the float flvm its alw~ for 
le remplacer.(Castorama 50.11) 
it to renew. 
(33) Tracer deux parall~.les h l'axe 
Druw two parallel-lines to the axi.~ 
vertical pour tapisser l:int~rieur du 
vertical for to paper the inside of/the 
cadre.(Castorama 70.04) 
fT'a~t e. 
While the intermediate steps between action and goal 
are obvious in example 32, it is far front self-evident 
in example 33 that the action enables the placing of 
the moulding that will frame the wall-paper when it is 
eventually hung. 
Interestingly. in half of its occurrences in Enablement 
expressions: "pour" precedes "l)OUvoir" (':to be able"), 
making the goal explicitly an enabling state and denying 
other expectations: 
(34) Diluer la peinture pour pouvoir la 
Dilute the paint for to be able it 
pulv~fiser.(Castorama 20.11) 
to spray. 
The unacceptability of the following example supports 
this analysis: 
(35) ".r.~ Flip the switch to be able to turn on the light. 
The most frequent reahsation within the sentence is 
with the conjunction "et" ("and"), which coincides with 
the English findings: 
(36) D6monter la buse et la nettoyer 
Dis-assemble the nozzle and it clean 
avec un petit pinceau.(Castorama20.11) 
with a brush small. 
Temporal markers of sequence account for the remain- 
ing 41(Yo of the examples. By definition, they cannot 
express Generation. In contrast to the English results, 
we noted a number of instances of "apr~s" (':after"), as 
well as "avant" (':before"): 
(37) D6monter le r6servoir apr~.s 
Dismantle the tank after 
d~viss~ les deux 
unfastened the two 
vis de fixation.(Castorama 20.11) 
fixing screws. 
avoir 
having 
The ordering of the action descriptions in the text 
is the reverse of that required in the world. The same 
is often true. of course, of expressions using "avant'. 
However, the latter appear calculated to preempt rather 
than provoke a rash action on the part of the reader. 
The Portuguese Results 
The Portuguese corpus discussed here comes from 8 
sets of instruction manuals written in Brazilian Por- 
tuguese. The mamlals are taken from consumer prod- 
ucts; they are aimed at a non-specialist audience and 
they vary from between 2 to 7 pages in length. As with 
our analysis of the French data, we set ourselves the 
task of examining the relationships between the Gener- 
ation and Enablement relations and syntactic forms in 
approximately 30 examples of each relation. In what 
follows, we present the result of our analysis of the 33 
Generation and 32 Enablement relations taken. A de- 
tailed accounting of the grammatical forms used to ex- 
press Generution relations is shown in table 5. Similar 
statistics for Enablement relations are shown in table 6. 
The results of these mappings, displayed graphically 
in figure 6: show a pattern of ahnost exclusive use of 
grammatical realizations for the two underlying rela- 
tions; we will argue that in fact there is no overlap in 
our data between expressions of the two relations. In 
general, the pattern of results for Portuguese is rather 
closer to our findings for French than for English. 
Expressing Generation in Portuguese In our Por- 
tuguese data; expressions of Generation most often in- 
volve the "para" connective, primarily followed by an 
infinitive (Para + Inf£n±tive) and on rarer occasions 
by a nominal phrase (Para Noun). Consider. for ex- 
ample; the following instructions from mammls for an 
electric saw and a mini vacuum cleaner: 
68 
7th International Generation Workshop • Kennebunkport, Maine • June 21-24, 1994 
F nction  ..... 
Figure 6: The Portuguese function to form mapping of Generation and Enablement Relations 
Grammatical Form Count .Percentage 
Para+Infinitive 23 70% 
Se+Subjunctive 3 9% 
Para+Phrases 2 6% 
Sequence 1 3% 
2 other forms 4 12% 
Total: 33 
Table 5: The frequency of Portuguese grammatical 
forms for Generation Relations 
Grammatical Form Count Percentage 
Sequence: 25 78% 
Depots Conditions 4 12% 
Antes Conditions 3 9% 
Total: 32 
Table 6: The frequency of Portuguese grammatical 
forms for Enablement Relations 
(38) Para colocar a serra na posi~$o de 
To put the saw in the position of 
corte obl~quo: solte a porca borboleta e 
cut obl.ique flee the scTvw butterfly and 
incline a sapata para o ~mgulo 
incline the shoe to the angle 
desejado.(Bla&&Decker) 
desired. 
(39) Para a montagem: encaixe os 
For" the mounting, fit the 
implementos conforme demonstra a 
implements consistent with what sho'ws the 
figura 07.(Dustbuster) 
figure 07. 
Of the remaining forms: two are a direct consequence 
of the preference fur using para forms: combined with 
the tendency in the language towards long sentences. A 
good example of this is shown in the following: taken 
from a manual for a garden tractor: 
(40) A CADA 100 HORAS: 
EveT'y 100 hour.% 
drene o resto de combust~vel quando 
drain the rest of the fuel when 
o n~vel estiver no mmmlo 
the level is at the minimum 
para retirar a ~igua e impurezas 
to get rid of the water and impurities 
contidas no tanque 
contained in the tank 
para n$o sobrecarregar o filtro 
to not overload the filter 
e dar nlelhor durabilidade na bomba 
and give better endurance in the pump 
injetora e bico injetor, 
injector and jet, 
resultando em consequ~ncia: em melhor 
~esulting in consequence, in better 
funcionamento do motor.(Tobatta) 
performance of the motor. 
Sentences such as these: with long chains of Genera- 
tion relations; are not unique in Brazilian Portuguese 4. 
As noted; the preferred signal of the Generation rela- 
tion is "para". For stylistic reasons (i.e.: to avoid using 
the same connective to link the four instances of Gen- 
eration within this sentence): the author has chosen to 
introduce some elegant variation. The third connective 
"e" (':and") appears as a Sequence: but what is in fact 
happening here is that the author has elided the "para" 
from the expression % para" ('Sand to"). Sequence is in 
fact reserved for Enablement~ but here the "e" (a typical 
example of a Sequence marker) is not misleading: since 
the reader has been primed by the 'rhythm: of the pre- 
ceding sequence of "para's to interpret it as yet another 
one. Similarly: the choice of the expression resultando 
em conseq~ncia seems to be motivated by the desire to 
avoid using yet another para. This expression of Gen- 
eration on its own would be extremely awkward: 
(41) ?? N~o sobrecarregue o filtro 
Do not overload the filter" 
resultando em conseqii6ncia em melhor 
resulting "in consequence in better 
funcionamento do motor. 
pe~forw~ance of the motor. 
4This example demonstrates the trazasitivity feature of 
Generation, discussed by Goldman (19701}. 
69 
7th International Generation Workshop • Kennebunkport, Maine • June 21-24, 1994 
Ex~pressing Enablement in Portuguese The 8 
Portuguese manuals we report on here contained many 
fewer instances of Enablement than Generation. Inter- 
estingly, although para is normaUy also the default ex- 
pression of Enablement, we found no instances of it in 
our data: presumably because this form is such a strong 
marker of Generation. Instead, we find Enablement to 
be expressed through signals of the temporal ordering 
of the actions involved. This we refer to as Sequence: 
a string of imperatives, either in consecutive sentences, 
or within a sentence and linked by the conjunction "e" 
("and"). 
(42) Recoloque o filtro de p6 e monte 
Re-put the filter of dust and mount 
novamente o aparelho.(Dustbnster) 
once again the appliance. 
or Antes/Depois Conditions, sequential steps signaled 
by the use of "antes de" (':before") or "depois de" (':af- 
ter") or their equivalents (e.g. "diante" ("before") or 
"ap6s" ("after")): 
(43) Antes de us~-lo leia atentamente as 
Before using it read carefully the 
ins t ruq.Ses.(Dustbuster) 
mstruct.i.or~ : 
(44) Ap6s colocar o plugue na tomada, 
After to put the plug in the outlet, 
pressione para baLxo a tecla do 
press downwaTds on the key of the 
interruptor.(Caf6 Quatro coffee machine) 
inte,~'uptor. 
As in the French examples: the use of expressions such 
as antes de tend to arise in cases where the writer ap- 
pears to be attempting to preempt rather than provoke 
a rash action, as in the following: 
Desligue a serra da tomada: antes de 
Unplug the saul from the outlet, heloT, 
fazer qualquer ajuste.(Black~Decker) 
to make whatever adjustment. 
Discussion 
Our study so far provides some important guidelines 
for representing and expressing Generation and Enable- 
ment: which we are soon to begin implementing in our 
system for generating multilingual drafts of instructions 
for software use and for carrying out administrative pro- 
cedures. The Generation and Enablement relations have 
proved here to be a valuable means for capturing, in a 
language-independent form: some of the deep semantic 
relationships underlying procedures. The data analysis 
we report on has suggested the range of appropriate syn- 
tactic expression of Generation and Enablement in En- 
glish: French and Portuguese, and regions of potential 
amlfiguity. For example, the imtications are that Pour 
in:finitive expressions in French shcmhl be avoided 
where possil)le (except where fldlowed by "pourvoir"); 
as should be expressions in English which use the To 
infinitive despite the fact that they are I.) 3, far the 
most frequent form of Generation in our English corpus. 
Clearly, the results we present here are only indicative; 
as we increase the size of our corpus we expect to gain a 
clearer picture of the relationships between the procedu- 
ral relations and their linguistic expressions. So far we 
have not examined the influence of the wider context in 
which the instances of Generation and Enablement oc- 
cur in the corpus, and although we are informed on the 
allowable and preferable mappings from Generation and 
Enablement to language, we are not yet in a position to 
identify the influences affecting the selection of the most 
appropriate expression for a given instance of Genera- 
tion or Enablement in a particular context. In addition, 
the rSle of RST in our final architecture is still not clear; 
we suspect that if a rSle exists, it will be to inform the 
choice of expression from within the allowable set. 
Acknowledgments 
\Ve are grateflfl to Lucia 1Rino for her comments on the 
Portuguese material and to our three annonymous re- 
viewers. 

References 
Balkanski, C. T. (1993). Actions~ Beliefs and Intention.* in 
Multi-Action Utterances. PhD thesis, Harvard Univer- 
sity. 
Delia, J., Scott, D., and Hartley, T. (1993). Knowledge, 
intention, rhetoric: Levels of variation in multilingual 
instructions. In Proceedings of the Workshop on Inten- 
tionality and Structure in Discourse Relations, Jtme 21, 
Columbus, OH, pages 7-10. 
Di Eugenio, B. (1992). Understanding natural language in- 
structions: The case of purpose clauses. In Proceedings of the Annual Meeting of Association \]br Computational 
Linguistics, Newark,'DE, pages 120-127. 
Fikes, R. E. and Nilsson, N. (1971). STRIPS: a new approach 
to the application of theorem proving to problem solv- 
ing. Artificial Intelligence, 2:189-208. 
Goldman, A. I. (1970). A Theory of Human Action. Prentice 
Hall, Englewood Cliffs, NJ. 
Mann, W. C. and Thompson, S. A. (1988). Rhetorical struc- 
ture theory: Toward a functional theory of text organi- 
zation. Text: An Interdisciplinary Journal\[or the Study 
of Text, 8(2). 
Pollack, M. E. (1986). Inferring Domain Plans in Question- 
Answering. PhD thesis, University of Pennsylvania. SRI Technical Report SRIN-403. 
RSsner, D. and Stede, M. (1992). Customizing RST for the 
automatic production of teclmical manuals. In Dale, 
R., Hovy, E., RSesner, D., and Stock, O., editors, A,- 
peers of Automated Natural Language Generation, Lec- ture Notes in Artificial Intelligence 587, pages 199-214. 
Springer Verlag, Berlin. 
Sacerdoti, E. D. (1977). A Structure \]br Plan,* and Behavior. 
Elsevier, New York. 
Vander Linden, K. (1993:). Speaking of Actions: Choo,~ing 
Rhetorical Status and Grammatical Form in Instruc- tional Text Generation. PhD thesis, University of Col- 
orado. Available as Technical Report CU-CS-654-93. 
