Sunday, June 24, 2012

When the Word Does Not (Appear to) Add Up

One day, gasoline was labeled "inflammable" and the next, "flammable"  and "extremely flammable."  Or so it seemed.

I was unaware of this change until it reached a critical mass. It may have seemed sudden, but the transition from inflammable to flammable took several decades. The Ngram, graphed below and discussed by The Grammarist, shows that flammable overtook inflammable somewhere around 1978, at least in the publications Google includes in their corpus.

The change occurred to avert possible misinterpretation of the danger signs. (Tap on the graph curve to show stats by year.)

So, inflammable is in transition, losing its prefix in-. In American-English, the transition is  nearly complete. However, the related derivations retain the prefix, as with inflammatory, inflammation, inflammability. This word is also related through the root to inflame, inflaming, and the French loan words flambeau, flambé and flamboyant. See more words from the FLAM root family at Word Info. (See comments for discussion of possible spelling of base as FLAMM.)

Misinterpretation occurred because of the prefix in-. The word inflammable is parsed into in- + flam + -able. The root FLAM denotes 'to kindle, to set on fire, to burn.' This word came to English via Latin by way of French. The prefix in- comes in several spelling forms: in- (indirect), im- (improper), il- (illegal), and ir- (irreligious). 

The prefix in- has three meanings or interpretations. The most commonly used sense conveys 'not' or 'non' as in insensitive, inedible, illegal, implausible, irreverent, etc. 

The secondary meaning for the prefix in- is  'into, in, on, upon' as in insert, incandescent, and implant.  

The third sense of the prefix is qualitative. Sometimes it serves as an intensifier, much like we use very or extremely. 

So, basing word meaning strictly on the meanings of each morpheme, inflammable could be interpreted in three possible ways:
  1. 'not flammable'
  2. 'flammable from within, able to burst into flame'
  3. 'extremely flammable' 
The second and perhaps even the third options are correct; inflammable indicates that something is "flammable from within, able to burst into flame, able to be set on fire." See Online Etymology Dictionary.

But apparently, and understandably, some folks thought inflammable meant 'not flammable' just as inedible means 'not edible.'  So, the decision was made to change the danger signs.  Gasoline and other combustibles are fairly consistently labeled FLAMMABLE. Today, FLAMMABLE and INFLAMMABLE are synonyms. They mean the same thing. They both warn of fire. But FLAMMABLE is seen far more frequently.

To make meaning quite clear, non-combustible materials that are not flammable are now termed non-flammable, thus avoiding the polysemous prefix in- completely. Why muddy the waters with the negative sense of the prefix in- when we have only just clarified things? And using green for the signage helps to convey the meaning. A safe color, green is go.

Because inflammable derives from Latin, it is likely to have some look-alikes or cognates across the Romance languages. But when it changed to flammable in English, did it also change in Spanish? In French? In Italian? Not so much. See signs.

To reduce confusion, perhaps we should just do away with words and communicate through pictures, icons, signs. This involves semiotics, the study of how signs, symbols, and icons are interpreted and used to convey meaning. A sort of "sign-ology" (not to be confused with scientology).

Before Bill Bryson kindled my interest in morphology and word origins with his engaging if not authoritative book, The Mother Tongue: English and How it Got That Way, I assumed inflammable was just a poorly constructed word. I did not realize that the prefix in- had more than one meaning; instead, I consistently interpreted the prefix to mean 'not'.

But take note! This confusion did NOT cause me to underestimate the danger, spelled out on red cans and orange signs: Given what I know about the world (ah, so we infer!) and particularly about gasoline, I decided I would simply have to ignore the pesky prefix and go with my gut. Those gas pumps are gonna readily burn even though the warning sign does not seem to indicate so.

Perhaps my own coming to terms with inflammable sheds a little light on how we all  process language. When it comes to words, one of the questions linguists debate is which is processed first or the fastest and/or to the greatest extent, the whole word or the morphemes that make up the word? Well, both, actually. In addition, it is context that floats our language boat. Thus, one of my favorite reads: In the Beginning was the Word (Aronoff, 2007) and, I dare to add,  "In the End, the Word" and "Along the Way, the Morphemes."

Implications for Teaching Vocabulary: Teach students whole words. Also, teach them morphology (prefixes especially). Teach them to pay close attention to context. Help them learn to trust their own gut--to develop self-efficacy when interpreting word meanings, by drawing clues from context and from morphemes, and from knowledge of the world. Also, teach them to use a dictionary when needed. This aligns with the Common Core State Standards for English Language Arts. Here is an example from grade 5:

COMMON CORE STANDARD: Vocabulary Acquisition and Use L.5.4.  
Determine or clarify the meaning of unknown and multiple-meaning words and phrases based on grade 5 reading and content, choosing flexibly from a range of strategies.
  • Use context (e.g., cause/effect relationships and comparisons in text) as a clue to the meaning of a word or phrase. 
  • Use common, grade-appropriate Greek and Latin affixes and roots as clues to the meaning of a word (e.g., photograph, photosynthesis). 
  •  Consult reference materials (e.g., dictionaries, glossaries, thesauruses), both print and digital, to find the pronunciation and determine or clarify the precise meaning of key words and phrases.

Also, encourage students to read, especially during the summer. Play word games (so many cool apps for word play--my newest fave is "4 Pics 1 Word"). Kindle interest in words, phrases, and language in general.

See also The Popular Prefix Survey Results


  1. Pat Johnson6/24/2012

    "The Mother Tongue: English and How it Got that Way" was written by Bill Bryson, not Bill Bryon. I've read several of his books, but hadn't discovered this one. Actually, I added several books to my Amazon wish list after I searched for this one! Thanks for the recommendation.

    1. Thanks, Pat. Yes, Bryson (typo corrected). Hope you enjoy The Mother Tongue. It may not be fully accurate, but it caught my attention.

      Let me know if you like any of the other books you found.

  2. Katy Mitchell6/25/2012

    It's 109 degrees here, Susan. The streets are flammable.

    We love Chicktionary! also Words with Friends and Boggle.

    I've been looking for a morphology app. Bluster Roots is better than Rootology, but I can't get my girls to play either of them. Do you know of a fun app for learning about roots and prefixes?

    YOU should make an app! =D

    1. Hi Katy,

      Funny thing. Someone recently pointed out Bluster Roots to me. I agree that it is more engaging than Rootology. We DO need some FUN morphology apps. Is anyone listening??

  3. Gail Venable6/26/2012

    Interesting post. There is another way that this word does not seem to add up. If the word comes from the verb "inflame", wouldn’t it be (in + flame + able), and if so, wouldn’t we expect "inflamable"? The OED actually cites uses of " inflamable" and "inflameable" through 1674, after which the "mm" spelling seems to have taken over. The Latin verb "flamm(are)" and the Latin noun "flamma" suggest the possibility of (in + flamm + able) as the structure. If we look at these words in their modern spelling, "flamm" seems to be the base of "flammable, flammability, inflammable" whereas "flame" seems to the base of "enflame, inflame, flaming, aflame, flameproof" .

    1. Hi Gail,

      You may be right.

      My logic for FLAM (rather than FLAMM) is as follows:

      Since Latin roots tend to be reductions of Latin words, I took the Latin noun flamma and the Latin verb flamm(are) and reduced them as much as possible, to FLAM.

      For spelling inflammable from the base FLAM, we apply the doubling rule: Double the final consonant if it is a lone consonant and add -able.

      flam + m + able >> flammable

      similar to:
      hum + m + able >> hummable
      stop + p + able >> stoppable

      This doubling rule will also work for inflammation, inflammatory.

      But, following your logic, it works to use FLAMM as the root. Question: Do we have other bases that end with two identical consonants? We have fluff, mill, miss, jazz. But are there any that end with a double m?

      It may be that we have two different (yet related) morphological families, each with a different base: FLAM (or indeed FLAMM) and FLAME. Yet, the two families (if there are two) would still be related to each other, through the root,and would share similarity in form and meaning that helps the word-learner process the word.

      Also, I had wondered if this is a kind of "twin root" as we see with VERT/VERSE (both Latin for "to turn or to change" as in reverse and revert or vertigo.

      Thanks for contributing. I look forward to hearing your thoughts.

  4. Nice discussion!

    This word comes up so often in my work because people often treat FLAMMABLE as evidence of inconsistency in spelling as they assume the base must be FLAME, which clearly does not work. (flame + able --> *flamable). I am so pleased, therefore, to see this investigation between two plausible structures that would both result in the spelling we have.

    My own initial assessment of the etymological evidence that Gail cites makes my working hypothesis that FLAMM is the base. The reason I wanted to share this conclusion is not to say that it is right, but to share with teachers how I draw conclusions when I encounter information that is beyond my current understanding -- something that happens a lot!!!

    As you explain, the structurally plausible hypothesis of a base FLAM assumes a root that is reduction of the Latin words. And with that you just exceeded my current understanding of Latin. You’ve suggested a process in Latin that I am now curious to learn more about, but until I do, I cannot use it to draw MY conclusion of the base of FLAMMABLE.

    So if I was in a classroom with children and I was sharing this story, I would be very happy to tell students that there is an interesting debate about the structure of FLAMMABLE. I would use word sums like you do to show that both FLAMM and FLAM are structurally possible, and I would show that most peoples’ first assumption FLAME is not! I would then explain that one of the hypotheses implies a process whereby Latin roots are often reduced from Latin words, but I don’t understand that process yet. So until we look farther at that to make sure we understand it, we will leave our working hypothesis that the current English base is FLAMM based on the evidence of the Latin "flamm(are)" "flamma" cited by Oxford. We can always reassess our current hypothesis if and when we better understand how Latin words and roots work.

    These kinds of issues of competing hypotheses, one of which invokes linguistic knowledge I don’t yet understand come up all the time. That’s why I like to be very explicit with students about how I try to deal with these situations. The principle I would post on a classroom wall might say something like “Don’t go deeper than you know!” I want to use our scientific study of words to teach principles of scientific inquiry.

    I would argue that accurate but incomplete analysis is less misleading than inaccurate analysis that errs by going to deep.

    So drawing from that principle I'll treat FLAMM as the base of FLAMMABLE until I have evidence I understand that allows me to analyze more deeply.

    The reason I go through all of that here, is that I find that when teachers understand this principle they can gain confidence in drawing their own conclusions, and they are much more confident in getting started analyzing words. When they hit words that make them wonder, they have a principled way to respond. It's fine to have a wall of unconfirmed hypotheses of analysis that are tempting, but awaiting convincing proof. In fact that's a great way to introduce young children to the process of doing science!

    Although I try to use this principle, I am no stranger to the other kind of mistake too! If you are curious, here’s one such example:

    1. Hi Pete,

      Thanks for chiming in. I always like to hear from you, but this time, the argument you make is not persuasive enough, so I will stick with FLAM until convinced that FLAMM is more correct.

      I did check with Word Info and, indeed, the root is listed as FLAM. I know this means little to you, because there is no discussion of HOW this was decided, but it is interesting, at least. Word Info is not concerned with creating a morphological matrix; it is more concerned with tracing words to their root source, even if several matrices could be formed from the derivatives.

      I inserted the Word Info link for FLAM into the body of the post, and see it here.

      I did check out your reasoning about major at your REAL SPELLERS link. Interesting!

      I added a note (see comments) into the body of the post, to alert others to the different views.


    2. Gail Venable7/01/2012

      Hi Susan,
      You could be right about the base of FLAMMABLE being FLAM rather than FLAMM. I don't have enough evidence to take a definite position on this. My main concern was to point out that the base could not be FLAME. And I agree absolutely that the families of words that come from FLAME and FLAMM (or FLAM) are etymologically related.

      As to whether there could be a base ending in MM, I'm not sure, but I'm not convinced that there couldn't be. This discussion has definitely motivated me to gather enough evidence to figure it out. The convention you mention above, about doubling FF. LL, SS, ZZ at the end of monosyllabic words is not relevant because it applies only to free bases; i.e., those that can stand alone. The base of FLAMMABLE, whether FLAM or FLAMM, is a bound base and therefore would not be governed by this convention. In fact, there are bound bases that end in a single L, just as you have suggested for FLAM. Consider the Latin verb PELLERE "push, drive" which gives us REPEL EXPEL and IMPEL. these words share the base element PEL.

      Thanks for the link to Word Info. As you say, it does not provide the information we need to figure out the orthographic base of FLAMMABLE, but it certainly contains some interesting clues about its etymological family that I will enjoy exploring.

    3. Hi Gail,

      You make several good points, especially the distinction between FLAM and FLAME. Same root, but different orthographic bases.

      Together with LEX, the final double MM, FF, LL, RR, etc. question has been nicely addressed. Thank you!

      Glad you like the Word Info Site. What a treasure trove!


  5. I'm not sure that the point is whether FLAM or FLAMM is "correct", since there is actual evidence for both (by evidence I mean words in the language, not just some resource or person who says one or the other). FLAMM is etymologically sounder, perhaps, but FLAM is morphologically just as elegant. Either spelling could account for all of the examples that anyone has found.

    I do think the point is the use of scientific inquiry, rather than assertions about what morphology 'tends to' do or not do. And I also think the point is the richness of the discussion and the clarity of distinctions between etymological roots and morphological base elements.

    Besides the free bases cited like JAZZ and MILL, we have bound bases like the TERR in TERROR and the HORR in HORROR that end in a doubled consonant, and plenty that end in SS -- like MISSile, PASSion, reGRESS. While HOR and TER would work also due to the doubling rule, the etymology would point us to RR rather than to R, as the consonant is double in the etymological root.

    We also have NN in the ANN / ENN of ANNual / biENNial, but the existence of doubled consonants elsewhere in other word families isn't really evidence of anything. Rather, we have to gather evidence of the surface realizations of a base in the word family itself, and look at what the etymology informs us of.

    Finally, FLAM(M) ad FLAME are not twin bases, because there was no historical twin in Latin. Twin bases alternate within the same word family based on discernible patterns, and both twins are traceable to distinct Latin patterns in the root. Just because two different bases derive from the same root does not make them twins: both PLEASE and PLAC(E) -- as in PLACid and PLACate -- are derived from Latin 'placere' but are not at all twin bases.

  6. Hi LEX (Linguistic Educator Exchange),

    Thanks for adding to the discussion.

    I had forgotten about HORR, TERR, etc. Right you are!

    Interesting discussion of twin roots. Thank you.

    So much to think about in all these comments from everyone!



  7. There's no such thing as a twin root. The literature is quite clear on the terminology. There are twin bases, which, like twin humans, share a common origin.

  8. Apparently, the literature is not quite clear. Else, how would I have heard / read the term "twin root"? In this field of study, terminology differs depending on whether I read from a linguist, a lexicographer, or an educational psychologist.

  9. Real Spelling7/01/2012

    LEX’s point about “the clarity of distinctions between etymological roots and morphological base elements” is fundamental.

    Much of current schooling literature persists in incoherently categorizing English base elements as ‘bases’ if they are “Anglo- Saxon” (a term that modern linguistics avoids as it is entirely ambiguous in its reference), but English base elements that are of Latin/Greek origin it calls ‘roots’.

    Thus what are stated to be “Anglo-Saxon” bases have historical ‘roots’ (often, but not only from Old English and Old Norse - two languages that many, but not all supposed ‘Anglo-Saxons’ spoke), while the industry’s Greco-Latin ‘roots’ themselves have ‘roots’, but of a different sort. Such ambivalent use of the same term ‘root’ for distinct concepts is incoherent and, as such, is unacceptable in scientific discourse.

    lt is, therefore, hardly surprising that conceptual confusion results from this overlapping of the terms ‘base’ and ‘root’.

    As long as this terminological disjuncture persists, the complementary interrelationship and distinction between orthographic morphology and orthographic etymology can only be mulled over with imprecision.

    The orthographic morphology of Modern English is synchronic; it concerns the elements (written morphemes) of the Modern English lexicon, the evidence for whose identity is precisely those English elements that are identifiable by rigorous analysis with the standard lexical algorithm, the word sum.

    The distinction between the orthographic ‘base’ and the orthographic ‘root’ is explained in this short tutorial film.

    There is also a more detailed explanation of the orthographic terms ‘base’, ‘root’ and ‘stem’ in this eBook.

  10. The words abHOR and deTER are counter-evidence to my suggestion, above, that HORR and TERR are base elements; I have Real Spelling to thank for that clarification.

    The terminology does indeed differ, which is one of the contributing factors to the deep word murk that lingers in language education. Encountering a term (or an explanation) in print doesn't make it accurate, of course -- een the OED lists *-tion as a suffix. Where terminology diverges, we ought to seek precision and clarity.

  11. Anonymous10/10/2017

    May I simply just say what a comfort to uncover somebody who
    truly understands what they're discussing on the
    web. You actually know how to bring an issue to light and make it important.
    A lot more people really need to read this and understand this side of your story.

    I was surprised you aren't more popular since you most certainly have the gift.

    1. Thank you! Glad you found it interesting.


Comments are published after they are reviewed, to ensure they are not SPAM.