I would not use indexicals or demonstractives in a statement alegged to be a fact or to represent a fact. I would use explicit indices, date, time, place.
Pardon me for being "picky", but I had a very good English teacher during my high school years.
Indirect quote: John said this pencil is long.
Direct quote: John said "This pencil is long".
The first we assume to be transparent, in that the indirect quote is about the pencil and a characteristic named "long".
The second we assume to be opague, in that the direct quote is about an utteance.
The technical difference is the presence or absence of explicit quotation marks around the utterance.
The sentence, "John said this pencil is long.", is not about what the sentence "John said 'This pencil is long'." is about.
Since we general semanticists are supposed to be particularly aware of using quotation marks, we ought to be a particular and precise about making sure we properly punctuate a direct and an indirect quote.
So, I "assume" that Ben left out the missing quotation marks inadvertently.
"John reported the length classification of this pencil as long using the previously provided definition of 'long'." is a statement of fact that is transparent about the pencil's length relative to the classification system used. It can be rechecked by another observer.
But without a prior classification system to "pin down" the "meaning" of the term 'long', it, can not serve as an explicit category in a classification system.
Prepare three small buckets of water, at three significantly different temperatures, "hot", "lukewarm", and "cold". Soak your two hands simultaneously, one in the "cold" bucket and one in the "warm" bucket long enough to habituate to the temperature. Then plunge both hands into the "lukewarm" bucket. Perhaps one will feel "long" and the other will feel "short". 
[Devise your own analogy.]
Ben previously wrote, I could make the second statement factual by saying "John says this pencil is long." This new statement doesn't say anything 'about' the pencil, but instead 'about' John (or 'about' what John said).
Ben claimed the new sentence to be factual, namely "John says this pencil is long.", which is an indirect quote - not a direct quote - of John's words, because, in Ben's very words, "This new statement doesn't say anything 'about' the pencil, but instead 'about' John (or 'about' what John said). Ben has equated an indirect quote with the explanation of a direct quote. A direct quote requires that the actual words spoken be used and quoted, So the sentence would have to be punctuated as: "John says, 'This pencil is long.'" It was not.
Ben later wrote, "John said this pencil is long." <-- This statement reports John's opinion.
"John said 'This pencil is long.'" <-- This statement reports what John specifically said (as well as his opinion).
Only half right. It uses invalid (unsane) logic.
"Quoted sentences" are opague; we can not validly infer that they accurately "mean" what we might assign to the quoted words. There are numerous examples in philosophical literature to show why this is so.
John saw Mary with Dick, who was not Mary's husband. John did not know this. John, assuming that Dick was Mary's husband, asked of another who Mary's husband was. The other said, "Mary's husband is Bob", where upon John said, "Bob is quite tall, isn't he.".
In John's sentences, John is talking about Bill, not Mary's husband Bob. The quotations are opague, so we may not infer that John's saying "this pencil is long", a direct quote, represents John's opinion, (although it might).
In probabalistic terms, direct quotations often do correlate with the equivalent indirect quote, so if we take a guess that the quoted sentence may be taken transparently, we will more likely be correct than wrong. But we cannot treat this as so "100 percent of the time", as the above example illustrates. "Sane" reasoning, using only valid logic, cautions us to be aware that treating opague contex as if it were transparent is a probabilistic correlation, not a "truth preserving" one. We may assume it's so in any given instance, but we must, like any other map-territory relationship, be prepared for this map to be wrong. Mary was not with her husband, although John assumed she was, and his conversation did not give him any contrary evidence, so the direct quote of his words were not about Bob. John's statement, "Bob is quite tall", was not about Bob, so it did not represent his opinion about Bob.
Ben,
You are correct. You have missed my point.
The distinction is between an indirect quote and a direct quote. You made an indirect quote, but you attributed to it the property of a direct quote.
You write Charles replies, "John said this pencil is long."
This sentence is a direct quote that contains an indirect quote.
What you wrote is a direct quote of what Charles said. Accepting the quote as meaning the conventional interpretation of the words in the quote of Charles is taking Charles words as "transparent", that is, that what they mean to us, the reader, is what they meant to Charles, the speaker. That is what "transparency" means. But direct quotes are not "transparent"; only "indirect" quotes are transparent.
The transparent phraseology would be Charles said that John said that the pencil was long.
Let me try another tack. I'll indicate levels of abstraction with distinct markers.
(Charles replies, ["John said {this pencil is long}."])
"{}" encloses an indirect quote. It has tranparency only one level up. Someone may abstract from the actual words John used and arrive at the words in the indirect quote. In other words, in an indirect quote of John, we may reasonably infer that the words used in the indirect quote paraphrase in some way what John said - that we can abstract the meaning of the paraphrase from the meaning of the words John used.
"[]" enclose a direct quote; we are not entitled to infer that what those words mean to Charles matches what they mean to us.
Philosophy has shown examples that disconfirm that theory. But, because people much more often than not stick with fairly conventional meanings, the probability is fairly high that the "transparent" interpretation is often the correct one. But logically it not a valid inference, because there are counter-examples; the "theory" that direct quotes may (always) be taken as transparent is disconfirmed.
"()" encloses your assertion, and it is a statement containing a direct quote.
If you had had "Charles replies, 'John said, "This pencil is long."'" you would have had a direct quote of a direct quote. It would be opague all the way down.
If you had had "Charles replied that John said that this pencil is long", you would have had an indirect quote of an indirect quote. It would be transparent all the way down. (Although, general semanticists would see a map of a map each made by different map makers, and would reasonably infer that this transparency might not be depended on.)
In your example, Charles was clearly sent to the dean for repeated tardiness. John's answer could indicate that he was mentally low functional and could only remember the last part of John's tirade, that is, assuming that Charles even got John's name right, especially since he had been using the same false excuse. It could also indicate that Charles knows full well why he is being sent, but he's "playing dumb" on purpose, so as not to be "putting himself on report".
In any event, you have provide a direct quote of Charles, and we cannot reliably conclude that what Charles actually said accurately reflects "the facts" of the matter. Such is the nature of a direct quote.
Indexicals are words like 'I', 'You', etc., which depend on context.
Demonstractives are words like 'this', 'that', etc; they also depend on context.
Ben, you wrote
Your "I could make the second statement factual" describes the characteristic of a "direct quote".
quote:I could make the second statement factual by saying "John says this pencil is long." This new statement doesn't say anything 'about' the pencil, but instead 'about' John (or 'about' what John said).
Your "by saying 'John says this pencil is long.'" is punctuated as an indirect quote because it does not have quotation marks around 'this pencil is long'.
You went on to say, "This new statement doesn't say anything 'about' the pencil, but instead 'about' John (or 'about' what John said).", and in doing so you are corroborating that you intended to write a direct quote.
You talked about a direct quote, but you punctuated the example as an indirect quote.
If you do not follow this, then perhaps you have some mislearned or misinformation as to the difference between direct and indirect quotation, of which there is plenty of information on the web.
Ben, see http://free-english-study.com/grammar/reported-speech.html
Ben,
Paraphrase is common, but it is not a required distinguishing characteristic for indirect quotes. The absence of quotation marks is.
You saw what you wanted to see. A direct quote must use quotation marks or be set off double indented (large quotes). Moreover an indirect quote of a simple assertion may have exactly the same words as corresponding direct quote, especially when pronouns were not used.
This is an indirect quote: John said this pencil is long.
This is a direct quote: John said, "This pencil is long." (The pencil is closer to John.)
This is a direct quote: John said, "That pencil is long." (The pencil further from John.)
Check the manual for writers and other such sources.
Moreover, your story implied an indirect quote, because, it "suggested" that he held up the pencil to be seen.
When the words are the same, dropping the quotation marks canges the "quote" from a direct quotation to an indirect quotation. The quotation marks are what inentifies in as a direct quote.
Paraphrase has the character of more or less, and the limiting case on the smallest side is zero, none.
If you think that it is "the same words" that make it a direct quote, you are misinformed. You may choose to be intensional and deny it, or you may choose to be extensional and check a number of sources.