Commentary

Faulty Punctuation Sends Reader up the Garden Path

When people read, they continually make predictions about the meanings and functions of words that they have not yet encountered. For example, when we read . . .

After the snow storm, the man shoveled X.

. . . we can predict that the function of “X” will be nominal (a noun that functions as the object of “shoveled”). Because of the context, we can make an informed prediction about the meaning of the subsequent words (in this case, “the driveway” or “his driveway”). After a reader recognizes part of a pattern, she assumes the rest. This chaining together of spontaneous assumptions is a significant component of the reading process and a major enabler of fluid reading.

You can think of the reading process as happening at different levels. First, there is spot reading. Spot reading does not fully engage the automated reading machinery. Instead, the reader assembles the words in short-term memory and transforms them into meaning by a much simpler process. In fact, many spot-reading tasks—such as reading a billboard or a PowerPoint slide—do not involve sentences at all.

Fluid reading, on the other hand, involves a highly complex, automated process with many moving parts [1]. The way in which we read is highly flexible and depends in part upon our purpose [2, 3]. McWhorter suggests three levels [4]:

  1. Reading for a Test: 150 to 250 words per minute
  2. Reading for Pleasure: 250 to 400 words per minute
  3. Reading for a specific fact (a.k.a. scanning): 600 words per minute or higher

Carver suggests five basic reading processes based on his research [5]:

  1. Memorizing: 138 words per minute
  2. Learning: 200 words per minute
  3. Reading for Comprehension (called rauding): 300 words per minute
  4. Skimming: 450 words per minute
  5. Scanning: 600 words per minute

Still others emphasize that the complexity of the text greatly influences the way that we read, from the speed of reading to the way our eyes dart around the page. And some reading strategies admit unconscious adaptations and even improvised behavior of the reader, especially during novel reading experiences (consider a technical novice attempting to assemble a fancy sound system via a difficult user manual). Technical material challenges the reader to “keep up” with the reading machinery—to stay within the range of “reading for comprehension.” The reader depends upon fluid reading to transform technical prose into concepts that can be integrated into the reader’s memory. But fluid reading is delicate and easily disrupted, especially when the reader’s predictions are violated.

What happens when the words that the reader encounters violate his predictions? The reader typically gets lost. This confusion in the middle of a sentence is called a garden path, because the author has led the reader up the garden path and abandoned him among words that seem irrelevant to the context (the reader actually creates the context through interpretation). Now, forward to two writing tips.

Use Punctuation to Join Independent Clauses

Here is a good example of a garden-path sentence caused by faulty punctuation:

These values are used to perform fault location and estimation errors are compared with those for the reference case.

Here, the reader can work under the assumption that fault location and estimation errors are objects of the verb perform. That is, the “values are used to perform” two things: fault location and estimation errors. However, once that assumption is made, the rest of the sentence is bewildering because the reader does not predict that a verb will follow those two objects.

There is a rule in English composition that two or more independent clauses (clauses that are able to stand alone) must be joined by a comma plus a coordinating conjunction, such as and (other compliant means of joining them include a semicolon or colon). In the current example, two independent clauses are joined by a coordinating conjunction alone—without a comma—and because there is no intervening punctuation, the material in the first independent clause sprawls into the next. Placing a comma after fault location clarifies the relationships between the words in the sentence and facilitates fluid reading.

Set Off Nonessential Material with Punctuation

Every sentence has at least one independent clause consisting of at least one subject and one verb. In some cases, there are additional clauses or phrases that are nonessential to the main clause. That is, they are included to supplement the meaning of something else in the sentence but not to complete it.

You can test the essentialness of an element by asking the question, “Is it required to complete the meaning of the thing to which it is connected (such as the main clause)?” If the answer is “yes,” then it is essential and can be conjoined to the material that it modifies without punctuation. (However, when an essential phrase or clause occupies the beginning of a sentence, a comma should follow it.) If the answer is “no,” then it should be joined to the sentence with punctuation, no matter where it resides in the sentence.

Here is another example of faulty punctuation that may lead the reader up the garden path:

For both methods, the data rate increased for all locations leading to improvement in accuracy.

The reader encounters the problem with this sentence at the word locations. What follows that word is certainly a phrase; specifically, it is a present participial phrase, which behaves like an adjective and is formed by adding ing to the base of a verb. In this case, the participial phrase “leading to improvement in accuracy” is nonessential because it is not required to complete the sense of the material preceding it. Without a comma to set it off from the rest of the sentence, the reader may mistake the phrase as essentially modifying locations. However, the entire participial phrase modifies everything that precedes it, and if a comma were placed before the phrase, the independent clause would not sprawl into it (as it does).

References

  1. B. Connatser. “A Phonological Reading Model for Technical Communicators.” Journal of Technical Writing and Communication 27(1). 1997. pp. 3-32.
  2. McDonald, A. S. 1963. “Flexibility in reading.” In Proceedings of the 8th International Reading Association Conf.
  3. O’Brien, J. A. 1921. Silent reading with special references to methods for developing speed. New York: Macmillan Company.
  4. McWhorter, K. 1992. College reading and study skills. New York: Harper Collins.
  5. Carver, R. P. 1990. Reading rate: A review of research and theory. San Diego: Academic Press.