Sunday, March 28, 2010

Academic Families of Words

If we want to promote comprehension we must strive to convey academic vocabulary in every subject area throughout the day, and at every grade level. The specific term academic word does not refer to domain-specific vocabulary like photosynthesis or decimal. An academic word is often the scholarly counterpart to a high-frequency common word. For example, the verb show is a common word but demonstrate is academic. Academic words are often abstract and slippery, not concrete. We encounter academic words less frequently but they are certainly not rare, and furthermore, they appear across subject areas, especially after third grade. Another example is shown in the picture to the left. The word find is academic as used here; students may know that find has the primary meaning of 'locate' but to carry out this task they must also understand its academic denotation. 

In daily conversation, educators can easily insert an academic word for a common word to help students become familiar with it. Slightly and gradually segue to an academic register when speaking and listening. This vocabulary transition might begin in kindergarten:
  • You may get obtain a pencil
  • It was a bad severe storm
  • Let's meet gather, assemble, convene or congregate in the library
Academic words abound in cross-curricular texts and assessments. If unknown, they present a tremendous obstacle to comprehension (RAND Reading Study Group, 2002). Even if students understand the subject matter they can be confounded by scholarly language, as humorously depicted above. Academic words often create a lexical bar, separating the word-savvy from the word-weary and disenfranchised (Corson, 1984). Teachers can help at-risk learners surmount the bar if they persevere together. It's a worthy goal but not quickly met. 

Academic Word List. Research in Applied Linguistics by Averil Coxhead led to the AWL (Academic Word List) of 570 words. These are the most frequently encountered academic words found in texts across disciplines of study in secondary and tertiary school. Each word represents a morphological family of words, so analysis includes analytical, analyze and analyst, for example. Conveying words in families clustered around the same root should promote morphological awareness as well as vocabulary. The AWL has ten sublists ordered by frequency; sublist 1 contains the academic words most frequently encountered. Word families from sublists 1- 5 should apply fairly well to grades 6-8 and sublists 6-10 should apply to high school and beyond. See the good doctor's tips for using the AWL. The AWL may be especially helpful for learning English as a second language because many of these words can be recognized as a Spanish-English cognate (sample cognates). Explore the Spanish Cognates Dictionary from Latin America Links or translate with the the multilingual Woxicon. 

More AWL resources: Visit the University of Nottingham for Sandra Haywood's useful AWL resources, including the AWL Gapmaker, the AWL Highlighter, and the Concordance. These three tools have potential utility for developing and assessing academic vocabulary. From Harvard University, explore Catherine Snow's Word Generation resources for developing academic language. Finally, explore Gerry Luton's AWL exercises and assessments through the the University of Victoria. All of these resources can potentially be used to good effect in a learning atmosphere that includes vocabulary-pertinent peer discussions as well as practice in correctly articulating longer words as modeled by the teacher.

Again, wait not for high waters. Use academic words in primary grade conversations and instructions. Prepare the soil, so to speak. If the children can understand the concept behind the word, they can understand the word. If they can understand make they can understand generate, produce, and create. By all means, speak UP.

Finally, April Fool's Day is just around the corner, but don't let anyone capitalize on your credulity. Don't credit everything you hear! See derivations constructed with the Latin root cred meaning 'to trust, to believe' (more derivations here, see all four pages). I promise, this is no fool's errand; the root cred is worth teaching because most of its derivations are, in fact, (you guessed it) academic words! In fact, Butler et al. (2004) determined that many of the academic words in fifth-grade content-domain texts are derivations, especially with respect to science.  

Click image 2x to enlarge.
PS. I have inserted page 59 from NTC's Dictionary of Latin and Greek Origins (1997, pp. 59-60).  NTC gives cred as the 'common root' of credere and spells it cred. At the bottom of the image I list additional sources naming cred as the root. Many sources, like NTC, include credo, creed, and credenza in the family of words denoting 'believe, trust.' However, perhaps the spelling should be crede (see comments).

  • Butler, F. A., Bailey, A. L., Stevens, R., Huang, B., & Lord, C. (2004). Academic English in fifth-grade mathematics, science, and social studies textbooks. CSE report 642. Center for the Study of Evaluation (CSE)/National Center for Research on Evaluation, Standards, and Student Testing (CRESST). Retrieved 7-2-09 from
  • Corson, D.J. (1984) The lexical bar: Lexical change from 12 to 15 years measured by social class, region and ethnicity. British Educational Research Journal 10(2), 115-133. 
  • RAND Reading Study Group. (2002). Reading for understanding: Towards an R & D program in reading comprehension. Santa Monica, CA: RAND.


  1. Lovely blog this week, Susan! You're calling for awareness, care and precision in the ways teachers use language in the classroom. The language we use in our teaching and training shapes our students' thinking about the very concepts we are teaching.

    This is so critical in language education, that metalanguage -- language for language teaching. It's a core piece of my current research -- we have much to learn as a field in this area. Thanks for sounding the alarm bells about the issue!

    I've been working with the AWL myself with two groups of adult, college/grad school-bound internationals. The published course materials don't do much for me, as they largely ignore morphological information and entirely ignore etymological information that allows students to learn how English works.

    I have supplemented the course materials with Word Sums and Word Matrices, some from Real Spelling and many others that I've created following the demonstrable principles of how real English spelling works. My students LOVE it.

    Rather than just learning a set of words on a list, my students are learning how words work. They collect and use evidence, not hunches and guesses, in their acquisition of English. They hypothesize about possible connections among words and have the deductive tools to research those hypotheses. I'll get my own blog and give the details there. :-)

    But here's an example: if we hypothesize that the base element of words like < credit > and
    < incredible > is < cred >, as you suggest, then we can investigate that hypothesis with a word sum. When we do, we find that

    < cred > + < it > = *< creddit >'

    because of the doubling rule. If we modify our hypothesis to consider < crede > as the actual base (rather than relying on surface appearances), then we find that our whole system of orthographic rules (like dropping a final non-syllabic < e > before a vowel suffix) and conventions holds up!

    < crede > + < it > -> < credit >

    < crede > + < ule > + < ous > -> < credulous >

    And we've learned not only a couple of words, and not only their actual base element (as shown by evidence rather than by a surface observation or a hunch), but we've also learned a consistent deductive methodology for explaining how the written word works.

    Keep up the good word!

  2. Gina’s demonstration that, by strict structural analysis, the base element of such spellings as < credulous > must be < crede > allows us to understand two further points, one general to orthography and the other specific to this word family.

    The general orthographic point is that the base element < crede > is bound (an element that is only found in association with at least one further element). Such bound base elements are very frequently found among those of Latin and Greek-origin, but not exclusively so. Even bases from Old English (the usual term in linguistics for what used to be called ‘Anglo-Saxon’) can be bound.

    A common example is the bound base < hap > that is found is such spellings as < mishap > < happen > < perhaps > < hapless > and the splendid compound < haphazard >.

    Bases, then, are often free, but not necessarily so as a good proportion of them are bound. Affixes are, though, by definition bound.

    The point specific to the family of the base < crede > is that this was formerly the spelling of what we now refer to as a ‘creed’ - “formal statement of a system of belief”.

    Christians, for instance, are particularly aware of The Apostles’ Creed whose initial word in the Latin origin is < credo > “I believe”. Thus a < credo >, a direct loan word from Latin, refers to someone’s “I believe”.

    Old English referred to this statement of belief as < creda > - the OE suffix < -a > formed a class of nouns. This suffix regularly became < e > in Middle English, so the word was written < crede >, the structural base that Gina’s analysis revealed.

    This spelling < crede > was that of the newly-invented printed books of Early Modern English. Here is a phrase from Caxton himself.

    “The three credes the whyche our moder holy chirche singeth”

    The first edition of the Anglican Church’s ‘Book of Common Prayer’ (1549) refers to “the Crede”.

    “After the Crede ended, shall folowe the Sermon…”

    A century later, however, the re-spelling < creed > appears, as for instance in the Shakespeare First Folio, and has remained standard ever since.

    So the base spelling < crede >, implied and analytically demonstrable from its modern derivations, has a pedigree that stretches back a millennium or more. Sadly, we might say, it has lost its former orthographic ‘freedom’ in favour of what, frankly, is an unnecessary re-spelling when it is used independently of any derivation.

    You are welcome to download a matrix on the bound base element < crede > from this link.

  3. Thank you, Real Spelling and Gina. I have inserted sources citing CRED as the root directly into my post. Perhaps, despite multiple sources, you are correct and it should be CREDE. I cannot say.

    I do not know of the suffix -it, as used in Gina's word sum cred + it and shown on the matrix. To my knowledge, it does not exist.

    Speaking of your matrix, if I were teaching vocabulary--the subject of this blog--I would include creed, credo, credenza in the matrix. I would arrange them off to the side, in a pull-out, or format them in italics, etc. so as not to confuse anyone about spelling, but I would keep them with the rest of the family.

    A key goal of lessons in morphawareness is to help children form networks of meaningfully related words and not necessarily by an analytical summing approach. In Aronoff and Fudeman (2005) I read of other, less strict approaches such as the more synthetic item-and-process view (which is tolerant of irregularities).

    Yes, indeed, the bound root of happy is hap. Glad you mention it! This one is interesting.

    PS. Comment revised--and I am still not sure what to make of CREDE -- but I will remain open.

  4. Searching through my correspondence of the last decade I came across a mail in which, like Susan, I declared that there is no suffix <-it >. I have no recollection, either, of having seen <-it > in any formal or suggested list of suffixes.

    However, neither declarations nor lists are validated by who wrote them, but by the testimony of the evidence.

    If there is a suffix <-it >, the evidence will be found in the English lexicon itself.

    Here is a preliminary evidence bank.

    < credit > (compare < credulous >)
    < audit > (compare < audible >)
    < deficit > (compare < deficient >)
    < implicit > (compare < implication >)
    < illicit > (compare < licentious >)
    < orbit > (clearly a derivation of the base < orb >)

    In the case of the pairing < credit > - < credulous >) < it > is replaced with the suffix <-ule >.

    In the case of the pairing < audit > - < audible> ) < it > is replaced with the suffix <-ible >.

    In the case of the pairing < deficit > - < deficient >) < it > is replaced with the connecting vowel letter < i > and suffix <-ent > (the bound base < fice > is also found in such words as < suffice >, < efficacious >, proficient, and in the compounds < edifice >, < orifice >, < sacrifice >, < certification > and < artifice >).

    In the case of the pairing < implicit > - < implication >) is replaced with the suffix <-ate >.

    In the case of the pairing < illicit > - < licentious >) < it > is replaced with the suffix <-ent >.

    I was clearly wrong to deny the existence of a suffix <-it >.

    Then, and as with all orthographic discoveries, once the evidence of the probability of a suffix <-it > was revealed by the evidence, other discoveries await us.

    Entertainingly, for instance, the etymon of the bound base of < decrepit > is the Latin verb < crep(are) > “make a noise, rattle crack”. In Latin, the adjective < decrepit(um) > “noiseless” was a metaphor for aged persons who, as our Lewis and Short Latin Dictionary puts it, “creep about like shadows”.

    Items that are < discrepant > denotationally make differing noises.

    The third grade student (for it is mainly through student, not teacher, observations that I learn most) who put me right on the existence of a suffix <-it > had been investigating < visit >. “When I visit someone,” she wrote to me, “I say that I am going to see them. We have been studying the twin base < vide / vise > so I wondered whether when we ‘visit’ someone it means that we are going to ‘see’ them.” She added that the dictionary seemed to be making that connection, but didn’t understand what was written, so asked for our help with the interpretation.

    She was right. The Latin stem < vis(um ) > “see(n)” with the ‘frequentative” verb-forming suffix <-it(are> ) does have the denotation of “frequenting someone by going to see them”. The orthographic structure of < visit > was, therefore, revealed as < vise/ + it >, and we had a little celebration to toast the epiphany of the suffix <-it >, as we do every time we discover from the evidence that I was mistaken about something.

    So, yes; the suffix <-it > makes its proud appearance in the matrix of the bound base element < crede >, and much of the credit goes to a third grader making a semantic connection that had previously escaped me.

    Real Spelling

  5. I think you're right on the money for teaching vocabulary with a matrix -- and we can tackle spelling at the same time, rather than using different frameworks for both. Including < creed >, < credo > and < credenza > -- the Italian word for 'credence' in the lesson is the way to go.

    Melvyn and I frequently discuss which suffixes we can include in a matrix, which ones are English and which aren't. I might be so bold as to include < crede > + < o > in the matrix with an annotation that < o > is only in borrowed words and is not a productive English suffix. But < crede > upholds our whole system of how the written word works in English.

    Regarding < it >: those letters can be a free base (from Old English), a pronoun, of course. They can be a bound base, as in < exit > and
    < transit >, from Latin _ire_, 'to go.' So < exit > is 'to go out' and < transit > is 'go across.' The word < ambit> is 'to go around.'

    The letters < it > can also be a suffix, as in the following words:

    < im > + < plice > + < it >
    < orb > + < it >
    < de > + < pose > + < it >
    < habe > + < it >

    (< habe > as a bound base is confirmed if we analyze the morphology of < habilitate >).

    I'm sure Melvyn can give more of the etymological information on the suffix < -it >.

    I've cited many of the same sources as you, Susan, in my teaching for years. But what I don't find in the NTC or other places listing < cred > is the *methodology*. What methodology did NTC use to analyze the morphological structure aside from bold-facing the common letters in each word? That's a surface observation, and it makes teachers and students consult and take sources at face value rather than interrogating them and investigating the language themselves. After all, the OED lists *< -tion > as a suffix. A theoretical framework is great, but should never supplant actual linguistic evidence.

    I appreciate your open-mindedness. As teachers, we can never be done learning!

  6. Interesting process and you may be right!

    As to the methodology others have used in naming common roots, I expect there WAS a methodology (but it may not have been based in spelling). English words are morphophonemic, so I would guess the methodology had to do with SOUNDS. In all the words with the letters CRED (or CREDE), the sound is KRED, not KREDE.

    I wonder if the suffix -it is largely unknown because it is not sufficiently productive. Rather, *it is most often merely a string of letters at the end of a word.

    If that is the case, then teaching the suffix -it is not of high priority and could be confusing to the developing reader who wants to APPLY the suffix to make meaning of many words (where it would not apply). It could be a great find for a more capable reader (as your third grader may have been)

    I went to MORE WORDS and found a total of 351 words ending with the letter string *it. (By comparison, I found 965 ending with letter string *ive as in give, mass-ive, creat-ive .)

    Most of the 351 clearly do not contain the suffix -it. Below is an excerpt (some may not be real words--to find out, one must click on the word's hyperlink). One could sort them:

    sit (not the suffix -it)
    tacit (probably the suffix -it)

    * probit
    * profit
    * prohibit
    * prosit
    * pulpit
    * pundit
    * pursuit
    * quit

    * skit
    * skylit
    * sleekit
    * slit
    * smit
    * snakebit
    * snit
    * snowsuit
    * sobeit
    * soffit
    * solicit
    * spacesuit
    * spait
    * spirit
    * spit
    * split

    PS-- Gina spoke of the doubling rule. I cannot follow the logic:

    For the VCV(silent e) pattern, in HIDE for example (drop e, no doubling the d, add suffix with vowel, we get hiding). Hiding is pronounced with a long vowel i. (There are many more examples where this pattern applies, including transforming eve to eventide).

    So, why, in the CREDE to credit transformation (drop silent e, adding suffix without doubling) is credit not pronounced with a long vowel /kreed-it/?

    credible should be pronounced /kreed--ible/ etc. etc.

  7. Good questions.

    In the words < credit> and < incredible > it's true that the sound is /kred/ with a short /e/. But not so in < credence > with a long /e/ or < credential > where it has a schwa (shewa).

    I wish I could use real phonetic symbols in here, alas!

    Anyhow, what we need is a base form that can accommodate ALL of the surface representations.

    The issue of when a vowel surfaces as short or long is complicated, not elucidated, by the whole syllable typing structure popular in reading education. But when we approach it morphologically, and especially when we approach it orthographically, < cred > doesn't hold up.

    Even if < it > isn't a very productive suffix, we can still illuminate the issue when we look at other, more productive suffixes:

    < cred > + < ible > -->
    *< creddible >

    We double the final consonant in a bound base the same as we do in a free base, as evidenced by < happy >.

    We see the same pattern with
    < phone > and < phonic > -- in one word, the vowel is short. In the other, the vowel is long. You may be less likely to suggest that
    < phonic > "should be" pronounced as "phone-ic." It's not.

    Morphemes retain their meaning and their spelling, but their pronunciation changes -- hence the morphophonemic characterization. This is a principle many are already familiar with.

    Questions like "why isn't it "creed-it" only arise when we assume (erroneously) that phonology is the primary motivator of the English writing system. By the same token, we'd have to ask, if the base were < cred >, then why isn't < credence > pronounced as KRED-ence?

    So, if the sources you cite used a methodology based on sound, it fails too. Thus far, I haven't found anything in these sources that indicates a methodology for ascertaining what a base actually is in a group of Modern English words, other than consulting other sources or observing surface patterns.

    Of course, there's always "street cred" -- which is a clip.

    Fun food for thought!


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