Much technical
prose is harder to read than it needs to be. It seems hard to readers
not
so much because the technical concepts are hard, but because writers
are
more focused on content than on readers’ needs. If writers can pay
attention
to certain principles of prose construction, readers are more apt to
find
technical prose coherent. Coherence is a quality of perception on the
part
of readers. Writers can only use techniques that promote cohesion and
hope
for the best.
(This web page owes much to Prof. G. Emlen Hall and his law-student editors of the Natural Resources Journal at the University of New Mexico's School of Law, who annually bring me out of retirement to try to explain syntactic strategies clearly and coherently, and to Prof. Bob Bond at New Mexico Tech, who used to ask me to explain them to the students in the Electrical Engineering Senior Seminar. Thus the preponderance of examples used in this presentation come from law and engineering.)
Clear prose should
evoke the response "I see what you're saying." Clarity is often
achieved by sharp editing or rewriting, which implies thinking about
the needs and backgrounds of readers. It can be compromised by jargon
(more about that later).
Cohesion can be organic or inorganic. As with broccoli, organic is better.
Coherence is not entirely within the control of writers and editors. Perhaps readers who don't find your prose coherent are inexperienced or lack your technical background. But they may also be misled by long, dense sentences and paragraphs that aren't organically cohesive.
Disclaimer: this
presentation can do little more than suggest a path
to further learning. What's here is distilled in part from a
semester-long course I used to teach, and even that was a mere
introduction to applied syntactic analysis.
Syntactic strategies use analytic techniques to understand the
structure of sentences. “Syntax” and “grammar” are more or less
interchangeable terms here, but “grammar” is perhaps more confusing
because it’s often understood as “the rules for writing well or
acceptably”--having good grammar, in this sense, is like having good
table manners: it’s a social grace.
Traditional grammars were based on Latin, an inflecting language. English was an inflecting language until the 11th century, but now it’s a word-order language. English is also phrasal--a German might use a jawbreaking compound where an English speaker uses a phrase. The recognition of this crucial fact about English--that it nests phrases inside phrases like Russian dolls--grows out of Noam Chomsky’s seminal 1957 work, Syntactic Structures, which first viewed English as a discrete combinatorial system.
Techniques
that promote cohesion depend on attention
to syntactic structure, or grammar. Sentence diagramming
wasn’t much fun in school, perhaps, and wasn’t seen as having any
practical
value. But in fact, I contend, the sort of sentence diagramming that
follows
the analytic methods of post-Chomskian transformational grammar is very
useful to technical writers and editors who need to write either
connected
discourse for paper publication or short chunks to be used in a
screen-read,
hypertext environment.
This web page will discuss three kinds of problems that
compromise readability and show how syntactic strategies can solve them:
- Jargon, aka technical terminology, which derives from extreme syntactic reduction
Sentences that are ambiguous or hard to parse
Paragraphs that lack cohesiveness
A discrete combinatorial system (DCS) combines a small,
fixed number of finite elements in infinite numbers of combinations.
For example, the English alphabet contains only 26 letters, from which
we can create hundreds of thousands of words. These letters can be used
in certain sequences but not in others. “Kcta” doesn’t work in English,
though it might in some other language.
Rules govern combinations all the way from the “atoms” of language that the alphabet represents poorly to the most convoluted and complicated sentences. As applied to grammar, a DCS describes the rules for creating, transforming, and embedding phrases, of which there are only four kinds--noun phrases, verb phrases, adjective phrases, and adverb phrases.
Rules are tacit in the sense that people use them without thinking when they speak or write. Rules reside in the brain in much the same way a Javascript resides in a web page: only run when needed. These are the real rules of language, not to be confused with “rules” invented by language mavens (e.g., “Don’t split infinitives.” “Always use a subject-case pronoun after a be verb.” “Never end a sentence with a preposition.”)
Sentences may follow the rules, as we’ll see later, and still be hard for readers to understand. Syntactic strategies are simply ways to figure out what’s going on in a sentence so that it can be reconstructed.
Some background: In the DCS of
English, we create mental trees. An English sentence (S) consists of a
noun phrase (NP) and a verb phrase (VP). But the subject noun phrase
may
contain a noun phrase followed by an adjective phrase, and that phrase
could contain a noun phrase that contains a relative clause, so the
reader
has to hold open the slot where the verb phrase goes until the
relations
among all the elements of the subject noun phrase have been established.
The six verb types: 1) Be, 2) linking, 3) intransitive, 4)transitive with one object, 5) transitive with direct and indirect objects, and 6) transitive with object complement.
Nine transformations: 1) compounding; 2) passive voice; reduction of clauses to 3) relative clauses, 4) phrases, 5) noun clauses, 6) appositives, 7) absolute phrases, 8) adverb clauses, 9) nonrestrictive relative clauses. (See Morenberg 1997, Bush & Campbell 1995.)
Nesting is easy to see in a sentence with a number of
relative clauses:“The cat that Bill owns ate the mouse that stole the cheese that was left on the counter in the kitchen that he recently had remodeled.”
The simplest form of embedding is compounding, stringing whole sentences or sentence parts together with conjunctions:
“The sea was calm but the boat sank anyway.”
“The ferocious tiger and the mongoose were cornered by the villagers.”
“The mayor viewed the report with alarm and immediately called his chief of staff.”
Reductions make for greater density: “The cat that Bill owns” becomes “Bill’s cat.”
More extreme reductions give us initialisms. A group in New Mexico that purports to act in the best interests of the public is commonly known as NMPIRG (New Mexico Public Interest Research Group, itself a reduction of "group that does research in the interests of the public in New Mexico").
These simple clauses
were all teased out of a single sentence. Try embedding them to form a
single sentence.
Here’s the original from which all those clauses were
teased out. The first three sentences became the initialism NGO ( a bit
of jargon), and the first four sentences became the sentence’s head
noun phrase, NGO water trusts.The fifth through eleventh sentences embed into a participial adjective phrase which itself contains an infinitive phrase that contains an adverbial prepositional phrase. The first and second groups of sentences constitute the finished sentence's noun phrase, or subject.
The sentences in the right column provide the verb phrase, or predicate, that in the embedded sentence starts at “provide.” Note that the reader has to hold the verb slot open before getting an answer to “NGO trusts do what?”
Note that two of the right column’s sentences reduce to “water right holders.” This is a noun string, another bit of jargon formation that we’ll get to in a moment.
I
said earlier that sentences arise from grammatical rules, seemingly
hardwired into our brains, that act like background subroutines. We
produce convoluted sentences quite naturally.Our listeners and readers have corresponding tacit subroutines for trying to make sense of our utterances. These work on a probabilistic principle, as we’ll see in the next examples.
Does
time fly as an arrow does? Do you time flies as you would time an
arrow? Do time flies like arrows just as fruit flies like plums?Chances are you saw the “flies as an arrow does” meaning first. Your mental parser read “time” as a noun phrase and expected a verb to follow. The second meaning is less probable; you have to parse “time” as an imperative verb--an order. Since “time” is most commonly a noun and less often a verb, “time” as an adjective describing a kind of fly is least probable.
The other two sentences cause most readers to laugh: because of the placement of the final phrases, there’s a discordance between what we think the sentences are supposed to mean and what they appear to mean.
These
sentences cause your mental parser to hit its reset button.You tend to parse “the man who hunts ducks” as a noun phrase which should properly be followed by a verb phrase. When it isn’t, you go back and reparse: “The man who hunts” becomes the subject noun phrase, and “ducks” (most often a noun), taken together with “out”, becomes the verb phrase.
The other sentences require similar reparsing. They lead you down the garden path--that is, astray.
Most words can have their basic form changed by add-ons. Some, notably so-called irregular verbs, change form altogether: think, thought, drink,drank, drunk.
The
reason
for words' changing form is to help them play a variety of
grammatical roles, as
dictated by the structural rules.
"Holmes deduced that Moriarty was hiding behind the falls. His deduction was correct. However, Holmes’ deductive powers were of no use. His deductivity could not prevent his being thrown over the edge."
There is, of course, a reason why the
“morphed” forms exist. It is that sentences must work together, and the
morphed forms help that process.
Though unmorphed words are easiest for readers to deal with,
constructing prose units longer than a sentence becomes an
engineering problem: in the tradeoff between clarity and cohesion,
cohesion is more important, and the morphed forms help promote cohesion.
Which
is
it--a convention of old guys that play tubas, or a convention of
folks who play old tubas? A proposal that comes from the
railroad, or from the railroad workers, or neither?
Noun clusters, or noun strings, are reductions from full clauses that show the relationship among the nouns. They are a common means of streamlining semantic units for everyday use. The problem is that they are inherently ambiguous: as in this example, they can be understood several ways depending on how the syntax is reconstructed.
Trying to clarify with hyphens isn’t much help because there are several possible ways to group these nouns into noun phrases.
This
practice
of stacking nouns can be illustrated by a selection from an article
from
a journal for members of IEEE. This particular journal is not a
research
journal but is geared toward keeping electrical engineers abreast of
new
technology. Its articles contain plenty of examples of noun stacking, a
practice that’s exceedingly common in professional life.
Jargon often looks like initialisms or stacks of nouns, but it can be analyzed as if it were formed by a process of clause reduction. The advantage of such jargon is that it helps insiders communicate efficiently. However, outsiders may be unable to get the meaning even if they are able to undo the clause reductions--primarily because there are many possible underlying clauses.
BGA, here, stands for ball-grid array. This is just a further reduced noun string.
That one short sentence contains this information, embedded and transformed:
A die is used in manufacturing microprocessors of the Pentium variety; an array, based on a printed circuit, carries a medium to high count of silicon leads and transforms from the silicon to an array of balls made of solder under the substrate through a microvia (an electrical path < 0.1 mm between layers of a circuit board) ; an interface attaches a semiconductor to the next level of inter-connection by putting conductive bumps on the face of the chip and mounting it face down.
The subject noun phrase is also a noun string, probably parsed as “local development of energy sources” or perhaps as “development of local energy sources.” “Energy sources” is a simple noun string, or nugget of jargon, within the longer noun string.
In the predicate verb phrase, “infrastructure options” is
another
nugget within a longer noun string, which might be parsed as
“quantitative measures for assessing infrastructure options.”
Realign
parts: In traditional grammar, we speak of sentences
as
having
subjects, verbs, and objects or complements. Some newer grammars use
the
terms agent (who or what did it), action (what was done), and goal (who
or what it was done to) to identify the main parts of a sentence,
regardless
of its transformations. Making the subject of the sentence the same as
its agent, action same as verb, and goal same as object usually makes
clearer
sentences.
Example: "Bill caught six largemouth bass" aligns subject, verb, and object with agent, action, and goal. In "Six largemouth bass were caught," the goal is the subject and the agent disappears or is relegated to a preposotional phrase ("by Bill"). In "The catch was six largemouth bass," the agent has disappeared and the action has become the subject.
Convert Nominalizations into Verbs: Nominalizations are nouns that came from verbs. Verbs make prose more active, so look for nominalizations that you can make back into verbs. Example: turn consideration (noun) into consider (verb). Thus "Three factors were given consideration" becomes "Three factors were considered."
Passive to Active: A passive is made by putting a part of the verb to be in front of another verb's past participle. (In the sentence you just read, is made is a passive.) Sentences with verbs in the passive voice sometimes become awkward, and because the goal or the receiver has become the grammatical subject, the reader can't tell who did what. Changing "Three factors were considered" to "we [or whoever] considered three factors" makes the agent clear.
Cut out Clutter: Certain phrases pop into writers' minds unless writers fence them out. Phrases such as "on the basis of," "in excess of"; nouns such as "factors"; weak verbs such as "occurs," "experiences": all these make the reader work harder. Overqualification can also clutter sentences: change "the data would seem, superficially, possibly to indicate that . . ." to "the data indicate . . . ." Indicate is inherently less definite than mean.
See G. Gopen & J. Swan, “The science of scientific writing.” American Scientist 78 (1990): 550-555.
There’s an unstated expectation among readers that a sentence names something familiar in the subject noun phrase and then goes on to add new information about it in the predicate or verb phrase. However, this expectation is routinely violated in technical prose. Gopen andSwan suggest that, from a reader’s standpoint, technical prose is more often written with new info on the left and familiar info on the right because it’s the new info that most interests the writer. However, this practice makes reading unnecessarily difficult.
Editors have to be of two minds. With one mind, they have to understand the writers’ jargon and more importantly their intentions. In the other mind, they have to think as a reader: Is there enough context to let a reader connect? What terms need explanation? What noun strings need reconstitution? What is new information to the reader? How did we get from the “here” of this sentence to the “there” of the next one?
When paragraphs seem scrambled--new information up front, context somewhere in the middle of the paragraph--it’s up to the editor to realign it. Put the contextualizing, familiar material first, then the new.
This sentence requires a reader’s mental parser to
keep incomplete branches in suspension until the verb phrase finally
appears. Note that the subject noun phrase ends just before “is very
wide, ” creating a left-branching sentence. This sentence also uses a few reduced clauses that show up as stacked nouns: “fabrication processes,” “microvia boards.” Further reduction could have made the sentence still harder to decipher: “The cellular phone microvia board material and fabrication process range is very wide.”
Here’s
a diagram
that shows how the subject noun phrase is constructed. Note that it
uses
the two noun strings (which can be analyzed as reductions of longer
phrases)
essentially as single nouns.
Note also that an infinitive phrase nests within a past participial phrase that nests within a noun phrase that is the object of a preposition, and that the prepositional phrase is bound to another noun phrase. The whole conglomeration constitutes the noun phrase that is the subject of "is very wide."
Here’s another sentence from the same article. This
sentence branches to the right. Its subject noun phrase is fairly
uncomplicated, while the predicate verb phrase is more intricately
nested. From a reader’s standpoint, this right-branching structure is
easier to handle because it follows readers’ normal expectation that
familiar information belongs in the left branch, the subject noun
phrase, and new information belongs on the right.This convention is routinely violated in technical prose. Gopen and Swan (writing in American Scientist, 1990) suggest that, from a reader’s standpoint, such prose is written with new info on the left and familiar info on the right because it’s the new info that most interests the writer. However, this practice makes reading unnecessarily difficult.
Visual
observation of the waterflood displacement mechanisms after
asphaltene precipitation are shown to be similar to the mechanisms of
displacement
of the unprecipitated crude oil.This sentence was brought to me a dozen years ago by a student who was interning as an editor with Tech’s petroleum-research institute. He found it baffling, and so did I at first. My intuition said there was something deeply wrong with the sentence. At the time I used a quick-and-dirty bracketing method to isolate the principal components of the sentence. Bracketing involves eliminating prepositional phrases, subordinate clauses, relative clauses, participial phrases as ineligible candidates for the sentence’s subject and verb. It turned out that the core of the sentence was “observation are shown.” This is not just a disagreement in number between subject and verb; there is also a semantic disagreement because “observation” and “mechanisms” aren’t really similar. Also, somebody’s observing and showing, but because of nominalization and passive voice we don’t know who. And it may not be important.
Years later, I tried to diagram the sentence, as shown in the next slide.
This diagram is hard
to see at this scale, but what it shows is that there’s an apparent
subject phrase labeled “NP-head” and another noun phrase labeled
“NP-True Subj” which in the diagram is stuck inside a prepositional
phrase. “After asphaltene precipitation” shows up in the middle of the
sentence as a sentence modifier, not attached to any other phrase. The
verb is passive, and unless an actor doing the showing is somehow
important, it could be deleted. Noticing that “mechanisms of
displacement” and “displacement mechanisms” are really the same thing,
we could put the sentence modifier first in the sentence and get rid of
some redundancy:After asphaltene precipitation, displacement mechanisms of the waterflood are similar to those of the unprecipitated crude oil.
Note that this clarifying magic can be pulled off by a syntactician without a degree in petroleum engineering.
Sentences in isolation, like my previous examples, don’t
offer enough context to help us discern meaning. Hence, we rely on
longer units--paragraphs--to help us get meanings across. Paragraphs
can specify, elaborate, enumerate, illustrate--just as our freshman
comp teachers always told us. But some paragraphs simply work better
than others.More effective paragraphs observe several principles that lead readers easily through them:
- Consistent topics--if the topic of the paragraph appears at the left of the sentences, the paragraph will seem more coherent
- Those changed-form words and "misaligned" sentence structures mentioned earlier can help keep to topic to the left
- Forecasters help readers develop a
mental picture by orienting them in space, time, and sequence
By
using Ford
or a pronoun equivalent, this paragraph keeps the focus where it should
be--by keeping "Ford" forward in all the sentences if not actually in
the
subject position. The first sentence treats "Ford" as new information
and
places it in the verb phrase or predicate (right side). All sentences
thereafter
treat "Ford" as the old information for this paragraph. The transition
from the first sentence to the second uses what I call the inchworm
principle.
The inchworm strategy works best as a transitional device
between paragraphs--the last noun phrase in paragraph 1 becomes the
topic of paragraph 2; the last noun phrase in the first sentence
becomes the topic in subsequent sentences. Within this example
paragraph, there is a surface cohesiveness but no real coherence.
The transformations that are sometimes the enemies of
clarity can become the friends of cohesion. In the second sentence of
these pairs, the first sentence’s verb--rose--becomes a noun. But it’s
more effective as the subject of the second sentence, creating a short
left branch that allows new information to occur in the right branch.The passive “was sparked by” of course lacks an attributor. One could avoid the passive by rewording: “This rise resulted from an increased demand for barrel hoops.”
The art of using forecasters was probably something you
learned in freshman English. They are useful, but they add only a
surface sort of cohesion--inorganic rather than organic. Organic
cohesion comes with the kinds of sentence realignments that produce
single-topic paragraphs--cohesion through structure.Logical Forecasters: also, because, for example, therefore, furthermore, in addition, however, in conclusion, not only/but also, on the other hand, similarly, etc.
Sequential Forecasters: first, second, third, next, etc.
Temporal Forecasters: after, before, finally, soon, next, when, then, etc.
Spatial Forecasters: above, below, beside, adjacent, next, under, over, etc.
Readers
have
to work less hard if each paragraph or chunk becomes a unit in which
the
second and subsequent sentences have a consistent, identifiable topic
in
the left branch. (The first sentence names the topic, which may occur
in
its right branch.)
Noun strings of four or more words may be
fine for
technical audiences, but nontechnical audiences may need to have them
unstrung
to show relationships and supply background knowledge or context.
The last point addresses a personal peeve.
Many of the ads I see for writers and editors specify that applicants
must know how to use Word, Publisher, Photoshop and other software. It
seems from the ads that editors' skills with language are relatively
unimportant. But it is these skills that create documents more likely
to satisfy clients and customers and thus to increase the prosperity of
the companies and agencies. For more on this question, see
Chapter One of How to Edit
Technical Documents.
Suggested Readings
Max Morenberg, Doing Grammar. 3rd ed. Oxford, 2002.
Morenberg’s book makes the basic transformations easy to understand.
Steven Pinker, The Language Instinct. Perennial Press, 2000, partic. Chs. 5 & 7.
Pinker’s book is fun to read; the two chapters mentioned here explain tree structures in terms of both word morphology and syntax.
Don Bush and Chuck Campbell, How to Edit Technical Documents. Oryx, 1995, partic. Ch. 8.
Bush & Campbell present syntactic strategies as part of the technical editor’s toolkit--the first book on technical editing to do so. Oryx Press was bought out by Greenwood Press, which has allowed the book to go out ofstock. Check the library or email Campbell.
Martha Kolln, Rhetorical Grammar: Grammatical Choices, Rhetorical Effects. 4th ed. Longman, 2002.
Martha Kolln’s title reminds us that grammatical tactics serve rhetorical ends.
Joseph Williams, Style: Ten Lessons in Clarity and Grace. 8th ed. Longman, 2004.
Williams continues his two-decade struggle to strike a balance between clarity and cohesion.
Kim S. Campbell, Coherence, Continuity, and Cohesion: Theoretical Foundations for Document Design. Lawrence E Erlbaum, 1995.
Kim Campbell’s book tries for a sort of unified field theory of coherence, both in prose and visual design.
Examples of Prose in Need of Revision
The presentations from which this web page
is derived contain a number of examples of prose that is extremely
dense with jargon. They also have examples of prose whose authors were
evidently so deadline-bound or so enamored of the authorial voice that
they forgot to notice that their paragraphs aren't cohesive and their
sentences nearly defy readers' attempts to understand them. If
you find that your organization is putting out reports that are nearly
unintelligible except to the people who wrote them, you might want to contact Campbell about doing a
presentation or seminar for your people. Seminars can be tailored to
the kinds of documents your organization produces, and examples drawn
from your own documents.
But there are plenty of examples of
infelicitous prose you can sharpen your skills on if you keep your eyes
open. Standards of editing in the publishing field have slipped badly
over the past couple of decades. There are a number of possible reasons
for this slippage we needn't go into here. My own practice when
in need of examples has been simply to go to the university libraries
and somewhat randomly select texts and learned journals from the
shelves.
If you do find yourself drawn toward trying
to straighten out others' sentences and paragraphs, a certain humility is in order. As Joseph Williams has
said, "You can't write a book too often; eventually you might get it
right." After eight editions, he would know. Most of us don't have the
luxury of multiple rewritings in a time-is-money work
environment. Balancing clarity and cohesion is a constant struggle. We
might call it "prose engineering" because as writers and editors we're
always trying for the best compromise. Or at least the best
compomise we can get under the circumstances. I encourage people
to use syntactic strategies on randomly found paragraphs: if you do so,
you will become a more skillful reader and a more self- and
others-aware writer.