Lófasz

Laryngeal features:

You already have some background knowledge of phonology and phonetics, and the difference between these
two approaches to the linguistic reality (sounds, that is). What does phonetics do?

 studies sound signals

Yes, phonetics deals with the 'hands on' of linguistics. It studies sound signals, that is speech as it is emitted
(spoken) by the speakers. You can hear it, analyse it, dissect it, measure it. What are the various types of
phonetics? More than one answer is possible.

 articulatory
 acoustic
 auditory

Which branch of phonetics can we deal with most easily? Which one is most easily observable?

 articulatory

Yes, articulation is possibly easiest to observe. You can almost feel your tongue touching your palate, and so on.
The most difficult side is auditory phonetics, because it deals with how the brain processes sound signals. This is
an amazing feat of the human brain, but it's difficult to approach as it is basically neuro-science. To see behind
the scenes we'd need some pretty sophisticated technology.

Acoustic phonetics is also hands on but to have a go at it you must have good recording and analysing
technology (slicing sounds, measuring their intensity, pitch and so on). In other words, acoustic phonetics looks
at the physical properties of sound waves such as frequencies. We'll not be doing this in this course. What is
more accessible is articulatory phonetics. This branch of phonetics deals with...

 how sounds are produced by the speech organs

Yes, to be able to discuss phonetics we must introduce appropriate terminology such as... (more than one is
possible).

 Place of articulation (POA)

 Laryngeal features (voicing, aspiration)
 Manner of articulation (MOA)

Yes, these three will be the dimensions along which consonants will be analysed. For vowels we will need a
slightly different set of dimensions. To pinpoint the exact position of a consonant you will have to think in terms
of these three dimensions.

What comes now will be a revision of what you learned in foundations of phonology. Of course we are not going
to have the time to revise it all, we'll be concentrating on those aspects that will be relevant to this course.

Please remember, now we are describing sounds along those three dimensions we mentioned above. We'll be
describing sounds now. You might we wondering why I'm being emphatic on this: well, phonology will deal
with something slightly different (but phonology will have to draw its inspiration from phonetics).

But, first things first. We need to dispel some misconceptions. What is the difference between <bat> and [pat].
It's not the types of brackets, it's a whole philosophy behind this. <XXX> is used to represent...

 letters

Yes, it is letters, not sounds. Not all languages have letters, there are hundreds of languages which have no
spelling systems, but their speakers can still produce sounds, and sounds can be described. This is how we try to
refer to sounds: [pat]. Remember that this is a huge abstraction: no <bat> will ever sound the same [pat]. There
will be as many [pat]'s as many times you pronounce it. Still there will be something underlyingly 'batty' about
the many [pat]'s, by which we mean that the brain will be able to discern what lexical item we have pronounced
based on the features of the sound signals that enter our ears and get transmitted by the ear's three bony
construction into the brain. Quite some chemistry. So, we have [pat]1, [pat]2, ..., [pat]n, all representing <bat>.
Spelling is not an object for phonetics. It's very useful though, but remember spelling is not linguistics. Later we
will introduce /pat/ as well, which may be exactly what the brain discerns as fundamental behind the very many
[pat]'s. Yes, this will be the phoneme. But let's not jump ahead of ourselves.

Pick off which of these statements are wrong from the point of view of linguistics.

 pronounce this letter for me

 spell this sound for me

Yes, you can see the fundamental mistake people make every day :) They try to pronounce letters, whereas what
they really do is pronounce sounds (and spell letters, if they come from a language which does happen to have a
spelling system). Feel free to explain this to all those who have not been initiated :)

Let's tackle the first dimension briefly, that of voicing. Voicing is produced by...

 the vocal fold/chords in the glottis

Yes, it is the vocal folds that produce the voicing. The folds are housed inside the glottis, inside a rather intricate
'box', known as the voice box. Of course, there would be no voicing without the air pumped out by the lungs, but
strictly speaking it is the voice box that produces voicing. If the voice box is removed (as it happens in some
surgical operations), voicing is no longer possible (although breathing will continue unhampered). Why is
voicing important? Well, it supplies the fundamental frequency to vowels and voiced consonants. Voicing is like
a murmur that language uses to make contrast among sounds. Some sounds will be voiceless, some voiced. As a
matter of fact, this is much more complicated than this. There are many different types of voicing. Let's mention
some:

-- active voicing

-- active devoicing

-- passive voicing

-- passive devoicing

-- spontaneous voicing

-- spontaneous devoicing

All this is regulated by the vocal folds. However, there is another laryngeal feature that languages can make use
of, and this is aspiration. Aspirated sounds can be contrasted with non-aspirated voiceless sounds. Do you know
such a language?

 chinese, german, english

Yes, very interesting. English word-initially distinguishes voiceless from voiceless aspirated stops:

<bat> vs <pat> = [pat] vs [phat]

This shouldn't be new to you. English does not have actively voiced sounds. Voiceless (or lenis) sounds can be
voiced between voiced sounds, but this is passive voicing (they simply continue the voicing of the sounds around
them). Let's see some examples:

<bat> [pat]

<pat> [phat]

<tab> [thap]

<tap> [thap]

<abbot> [abət]

Can you see that English uses <b> (a spelling device!) to show voiceless (lenis) sounds? So <b> will be found as
either [b] or [p].
Voiceless aspirated sounds are known as fortis sounds. English spells them as <p>, <t>, <k> (as far as the stops
go). Note that a fortis sound can be either aspirated voiceless (<pat>) or non-aspirated voiceless (<bat>). And
lenis sounds can also become voiceless word-finally (<tab>). You can ask why don't <tab> and <tap> sound the
same? Well, vowels are slightly longer before lenis sounds, so strictly speaking <tab> is [tha:p], as opposed to
<tap> [thap]. How do we distinguish them? Thank your advanced brain for this, a unique construction in all the
universe.

This is a tremendously interesting subject, but we will only scratch the surface here. English, as we saw, has
contrasting fortis stops and lenis stops. Fortis stops are aspirated word-initially and before stressed vowels
(discussed later in the course), but they are devoiced word-finally, lenis stops are voiceless (non-aspirated) word-
initially and word-finally, but passively voiced between voiced sounds. A lovely division of labour.

Another class (in addition to the stops discussed above) where the lenis vs fortis distinction can be observed is
the fricatives: f θ s ʃ (fortis) vs v ð z ʒ. In the fricatives aspiration is difficult to observe (but there are tests,
discussed later), and there is a lot of controversy surrounding them (but this will not be discussed here). Observe
the examples: fine vs vine, thigh vs thy, etc. Lenis fricatives are devoiced word-finally, but the vowel before
them stays long: have [ha:f], live [lɪ:f], etc.

The distinction can also be observed in the class of affricates: tʃ vs dʒ (as found in church and judge).

Up to now we have seen how English can use aspiration and voicelessness (absence of voice) to differentiate
words: <pat> vs <bat>: [phat] vs [pat]. Word-finally there is no aspiration and no voicing either (as a matter of
fact, active voicing is not used by English). The contrasts come in pairs: [t] vs [th], [p] vs[ph], [k] vs [kh].

But is there any voicing that En uses? Yes, it's known as spontaneous voicing. The aerodynamics is quite
complicated but there are sounds for which voicing comes 'naturally', with no effort. These sounds are
spontaneously voiced (and do not have voiceless paris). These are the vowels and the sonorant consonants: the
interesting fact is that spontaneous voicing is stable (so there is no word-final devoicing, for example). Vowels
you know. The sonorants are: m, n, ŋ, r, l, j, w. They are spontaneously voiced and cannot be devoiced: <man>
is [man], not [man̥ ] (with the subscript circle showing devoicing).

So, in En we have fortis and lenis consonants, and sonorants (these are spontaneously voiced). In other words,
active voicing is not utilised by En (as opposed to Hungarian, for example). Vowels are also spontaneously
voiced. So, the sonorants are quite similar to vowels (but they are still consonantal, of course, for which we'll see
some proof later). Some call them hermaphrodites because they are part vowel, part consonant.

Now answer a few questions.

What consonants do we have in kick?

 [khik]

What is the best representation of big?

 [pi:k]

What is the difference phonetically between pick and pig?

 the length of the vowel

Yes, before word-final devoiced lenis consonants vowels are longer, but not before fortis ones (that lose their
asiration here), so [pɪ:k] for big, [phɪk] for pick. Actually, it is not only a matter of vowels getting longer before
lenis consonants, it is also about vowels getting shorter (or clipped) before fortis vowels. So in pick the short
vowel is actually 'extra' short, shown usually as [phɪ̆ k] (the breve shows shortness). This is known as pre-fortis
clipping. Note that long vowels are also shortened before word-final fortis consonants: pierce is [phɪ̆ :s],
diphthongs are also shortened, so cape [khɛ̆ jp]. Note that sonorant consonants are also shortened before fortis
consonants, so sent [sɛ̆ n̆ t] (vs send [sɛnt]). This is a new world, take some time to devour it.

So what is the difference between mend and meant?

 Pre-fortis clipping affecting [ɛn]

So, why do dictionaries not show this difference? They keep representing mend and meant as [mɛnd] vs [mɛnt],
rather than [mɛnt] vs [mɛ̆ n̆ t]. Because it is...
  predictable

Complexity is a matter of taste, what counts ac complex depends on the point of view of the one observing the
world. What is complex for us was not complex for Einstein. Dictionaries do not present predictable features,
such as devoicing, aspiration, pre-fortis clipping, etc. Dictionaries are not treasure troves of sound sequences put
on paper, they show the system behind the sounds. But this will soon usher in a new perspective on how a more
abstract system than that aiming to describe the sound waves can be implemented (but this is already phonology,
not phonetics, as you may have guessed). So what have we gathered so far? Choose the right answers below.

 Fortis consonants can be aspirated

 In En there is no active voicing
 In English <b> can represent a phonetically voiceless sound word-finally and word-initially
 Vowels are shorter (clipped) before fortis sounds
 Lenis consonants are passively voiced between voiced sounds (e.g., about)
 Vowels and sonorants are spontaneously voiced sounds
 Lenis stops are devoiced word-initially
 [XXX] shows pronunciation
 If you hear a voiced sound at the beginning of an En word, it can only be a sonorant or a vowel
 If you hear a word-final [t] it can be either fortis or lenis
 Vowels keep their length before lenis sounds
 The consonants in En are either spontaneously voiced or have no active voice feature

Note that the consonants in bend and bent are phonetically identical ([t]), one is a devoiced lenis C, the other is
fortis (which can never be voiced). We know that the two C's are not identical (in bend the vowel is longer,
in bent it is shorter). What is the name for this phenomenon, when a number of distinct entities have the same
pronunciation?
 merger

So, we have seen that Cs can be spontaneously voiced (these are known as sonorants). The other
(complementary) class is obstruents. They can be either fortis or lenis. Fortis sounds are voiceless and aspirated
(but may lose aspiration), lenis sounds are voiceless (but may be passively voiced between voiced sounds). This
distinction is best observed in the class of stops (as explained above: [ph] vs [p]).

Now we have two huge classes of consonants: obstruents and sonorants. But this is not enough for our
classification.

Which example contains lenis obstruents only?

 abode

As we saw in class, Hungarian has active voicing meaning voice can spread backwards (known as regressive
voicing assimilation):

(i) pap ban [pɒbbɒn] 'in the priest' (p becoming voiced)

(ii) mész ben [me:zbɛn] 'in the lime' (s becoming z)

Why is there no voicing assimilation in English backbone [pakpəwn] affecting <ck>?

 English has no active voicing

Very good, En has no active voicing (as opposed to Hungarian). Yet, in both En and Hu there is no voicing
assimilation in the examples below:

(i) Hu lapra [lɒprɒ] 'onto the sheet'

(ii) back row [pakrəw]

Why not?

 spontaneous voicing doesn't spread in either En or Hu

Manner of articulation (MOA):

We have seen that as far as laryngeal features are considered En does not have active voice in consonants:
sonorants are spontaneously voiced, obstruents are fortis (which can also be aspirated) or lenis (and lenis
consonants can be passively voiced between voiced sounds). We have these classes now:

-- obstruents:

A: FORTIS: ph (p), th (t), kh (k), f, θ, s, ʃ (basically the voiceless sound that can also be aspirated)

B: LENIS: p (b), t (d), k (g), v, ð, z, ʒ (basically the voiceless sounds that can be passively voiced)

(NOTE: the fricatives are an interesting class, we won't say too much about them now. It is possible that
fricatives are distinguished by voicing)

For better 'visibility' we will continue to show the obstruents as shown below (but remember that voice is not an
active feature of English):

A: FORTIS: p, t, k, f, θ, s, ʃ

B: LENIS: b, d, g, v, ð, z, ʒ

-- sonorants: m, n, ŋ, l, r, j, w (all spontaneously voiced)

This can't be the whole picture as we obviously need some further dimensions to
differentiate m from b from p from w. This is manner of articulation, which can be thought of as a cline/scale
along totally obstructed airflow to totally non-obstructed airflow. Of course, life isn't straightforward, neither is
MOA: there are a number of different types of manner of how a sound can be produced. Totally obstructed
sounds are the stops (the airflow encounters an obstacle in the mouth and it must be overcome for the sound to
be realised). The other other extreme is no obstruction. These are the vowels.

Between these two extremes we have a number of other manners. If obstruction is not complete, and there is a
narrow slit through which air escapes to produce friction, the thus produced are known as fricatives. If you
combine a stop and a fricative, you end up with an affricate, although they are a bit of a mystery. These three
makes up the class of obstruents (= they are produced with obstruction of air of some degree). These sounds can
be fortis or lenis (so the dimension of laryngeal features combines with the dimension of MOA).

The next huge class is that of spontaneously voiced sounds. They also come in a number of flavours. If there is
obstruction in the mouth, but not the nose, we have the nasals. If there is slight obstrution and it happens with
the sides or the tip of the tongue, the sound will be a liquid. Liquids are subdivided into r-type sounds (rhotics)
and laterals (l-type sounds). And one step before the vowels we have those sounds that are basically vowels but
because of their position they are best analysed as consonants: the semi-vowels or glides. This should be familiar
grounds.

Find the class of obstruents below:

 p t k b d g f, θ, s, ʃ, v, ð, z, ʒ

We have an abstract feature that cuts across the class of obstruent and sonorants from the poitn of view of
whether there is obstruction in the mouth or not. If there is, such sounds are known as non-continuants, if the
flow of air is unobstructed in the mouth such sounds are known as continuants. See below if you understand this.

+continuant -continuant

fricatives (obst) stops (obst)

liquids (son) nasals (son)

 glides (son)

Yes, the nasals are stops (there is obstruction in the mouth). Note also that the feature +/- continuant cuts across
the class of obstruents and sonorants.

The liquids and the glides are known as approximants (= sonorant continuants). We'll see where this will come
in handy.

Now, find the class that is defined as 'spontaneously voiced obstruents'

 there is no such class

Very good. Find the next class: 'fortis continuant obstruents'

 f, θ, s, ʃ

Find the 'spontaneously voiced continuants'

 l,r,j,w

The nasals are non-continuants!

Find the 'lenis non-continuants"

 bdg

Find the 'fortis continuants'.

 f, θ, s, ʃ

What is the name of such classes that are defined using a feature or features common to all such members?

 natural classes

What's the use of natural classes?

 Members of such a class will behave in the same way

Which natural class have we seen already?

 Lenis stops are voiced between voiced sounds

Find another natural class below.

 b d g p t k m n ŋ = non-continuants

Note that p t k s is not a complete natural class (f, θ, ʃ are missing). A natural class must contain ALL the
relevant members.

Find a complete class below.

 ptkbdg

p t k b d g = stops

p b t d (not a complete class of stops, or obstruents)

m n l r j w (not a complete class of sonorants = ŋ is missing)

j w l (not a complete class of sonorants, a number of consonants missing)

Try seeing this as a matrix, because there's a lot more coming our way.
Have another go, find a natural class below (remember: a natural class contains all the sounds that can be
characterized with a bundle of features that exclusively apply to them alone). Find the class of 'fortis continuant
obstruents'.

 fsθʃ

Place of articulation (POA)

Note that what you see in this scheme are the positions of the sounds according to their articulation. I'm sure you
see where I'm going with this: the phonetic places of articulation will not always coincide with the organisation
that phonology sees in them. Phonology will sometimes clump together two or more phonetic POA's into one.
But this is what can be expected: phonetics sees one aspect, phonology another. They complement each other in
a way that allows us to see facts from a number of perspectives.

Phonetics sees sounds in a piecemeal fashion, with all the details that phonology may regard as unnecessarily
detailed. What we mean by this is that phonology may simply collapse a number of POA's into a more general
one (we will see such examples in this course). Phonetics is overly rich in detail, whereas phonology is abstract
showing the barest minimum that needs to be shown (on the basis of contrast, not the richness of the sound
signals that come out of our mouths).

Look at the chart. You will see that as far as phonetics goes we have bi-labial (p b m w) and labio-dental (f v)
sounds. These can be called labials.

The next place of articulation is the dental (θ ð), the tip of the tongue touching the back of the upper teeth. The
next place is the alveolar, with the tip of the tongue touching the alveolar ridge (t d s z n l). As we move on, we
encounter the pre-palatal (or post-alveolar) POA (tʃ dʒ ʃ ʒ r). As we move farther down our mouths, we have
the palatal place, with j 'yod' being the only real palatal sound in English as far as phonetics goes. Other
languages like Hungarian have an array of consonants here (c 'ty', ɟ 'gy', ɲ 'ny'). We will see in this course that
English phonology considers the pre-palatal/post-alveolars and the palatals as
(simply) palatals phonologically. How do we know that phonology treats them as the same class?

 they behave in the same way

Yes, it is the behaviour of sounds that reveals to us what category they belong to phonologically. To be able
to do this, we have to observe processes. It is these processes that tell us what phonology sees as a natural
class. Phonetics is too rich to give us any answer in this field: ʃ ʒ, for example, are not real palatals. We will
see some phonological processes in this course that will reveal what phonology sees as behaving in the same
way. Establishing this will require some phonological analytic skills. We will be working on this.

Now, another huge general class of sounds is referred to as coronal (produced with the tip and/or the sides
(= corona) of the tongue). For some analysts, anything between the labial and velar places is coronal, so this
huge POA includes the dentals, the alveolars, the post-alveolars/pre-patalals and palatals. How do we
know they must be clumped together into this huge category? Well, yes, because there are processes that
show that they DO behave in the same way from one aspect or another. Where can we see this? We can't
just jump into this, we have to tread carefully. To see why the coronals are a uniform class we must be in
possession of a number of varied aspects of analysis. This will probably not happen in this semester, but
good things come to those who wait (and possibly take a specialised course in linguistics).

As already hinted at, the next POA is the velar (k g ŋ ɫ, w). Yes, w is here too: it is a labio-velar glide. It has
two POA's. We will see how we can decide which of the two places is its primary POA (the labial or the
velar). How? Through its distribution (= the places where it can be found as opposed to those where it
cannot). We'll see that distribution (similarly to phonological processes) will be able to reveal to us what
something is like and where it sits in our classification. Distribution and phonological processes single out
natural classes. You can say it's all about natural classes: a sound belongs to a given class on the basis of
how it behaves. It's the two sides of the same coin.

After the velar we have the glottal place (h ʔ). ʔ is a sound of English, of course, but is not usually found in
the chart of phonemes. Why? Yes, because it is predictable (it's a realisation of /t/).

Now we have revised the POA dimension. Note that the POA combines with the two other dimensions
(laryngeal features and MOA). Let's see if you can see the world of sounds in as the combination of these
three dimensions.
 Find the spontaneously voiced labial non-continuant(s)

 m

Find the fortis coronal continuant(s).

 θsʃ
Find the lenis pre-palatal continuant(s).

 ʒ
Find the spontaneously voiced alveolar continuant sonorant(s).
 l
Find the lenis bi-labial non-continuant(s).
 b

Find the spontaneously voiced velar continuant(s).

 ɫ

Find the fortis non-labial non-continuant(s).

 tk
Find the sonorant non-continuant(s).
 mnŋ
Which of these classes is spontaneously voiced?

 sonorants

Which of these classes refers to a MOA?

 continuant

Which of these is NOT a POA?

 sonorant
Which of these is non-continuant (= stop)?

 nasals

Which of these is continuant?

 fricatives

Which of these does not exist?

 nasal fricative

Which sound is the only real palatal of English phonetically?

 j

Which of these is not palatal phonetically, but behaves like a palatal phonologically?

 l

Middle English (Morphology, Syntax, Metre)

ou may ask how on earth do we know what Chaucer's pronunciation was? Shakespeare said that Chaucer was a
sloppy poet, but why? Not because of his nonchalant use of poetic imagery, but because of failing to meet the
requirements of the iambic pentameter. Yes, this is linguistics meeting poetry. What is the iambic pentameter?
(more than one answer my be correct)
  An iamb is a beat consisting of a stressed and an unstressed syllable (usually shown as X /)
 A poetic tradition introduced from the continent (it has no OE basis)
 A line having 10 or 11 syllables

Very good. An iambic pentameter has 10 or 11 syllables arranged int iambs [ɑ́ jam]. An iamb consists of an
unstressed syllable followed by a stressed syllable (schematically X /), so a line consists of 5 * X /. If the line
ends in a stressed syllable, the line is known as masculine, if it ends in an unstressed syllable it is a feminine line.
Chaucer's lines (with the exception of some, of course) conform to this underlying pattern perfectly.

Note: in English the pentameter involves a strict alternation of unstressed and stressed syllables, NOT short (=
light) and long (= heavy) syllables that form the foundation of the pentameter in other IE languages (Latin,
Greek, the Slavonic languages) and Hungarian as well. This is an interesting difference in how the ancient poetic
metre was adopted (and adapted) in English when this tradition was brought over from the continent.

Why did then Shakespeare say what he did?

 By Sh's times all word final schwa's were lost, so words generally end up shorter than they were
 Chaucer spoke ME
This is fascinating. Chaucer's words have schwa's that were lost in the transition between ME and early MoE
(eMoE) of Shakespeare. Plus some words were stressed differently. You will see that in our reconstructed lines
there are 10 or 11 syllables. A line starts with an unstressed syllable and is followed by a stressed syllable (this is
one iamb) followed by another 4 such beats (and potentially an unstressed syllable), giving you this beating
pattern of di da di da di da di da di da (di). What's more Chaucer wrote in couplets: the last word of line a
rhymed with the last word of line b, and so on. You can call this line final rhyming (which is very dissimilar to
alliteration known from OE poetry, that is internal rhymes based on the consonants of stressed syllables).

Can you see that the first beat (= foot in phonological terms) is defective, it lacks an unstressed syllable. Instead
it begins with a stressed one. This was employed by Chaucer (and many others before and after him) at the
beginning of poems to 'launch' the first line with a beat/stress, rather than an unstressed syllable. The rest scans
as it should (pronounced schwas underlined to avoid confusion):

Whán that Áprill wíth his shóures sóote

The dróughte of Márch had pérced tó the róote

You may object that with and to wouldn't normally be stressed in spoken English (being function words). This is
true, but in an iambic line they can receive this 'strengthening', so in a sequence of a stressed syllable followed
by two unstressed syllable (/ X X) the last one gets elevated (/ X /), known as dactylic strengthening for those
who are interested (Áprill with --> Áprill wíth). This often happens in poetry. We will not be involved with
scanning intimately, it deserves a course of its own.

Let's translate:

when April with his sweet showers

the drought of March had pierced to the root

(NB: drought = a period of dry weather)

Why does Chaucer invert soote shoures to shoures soote? This is not a typical order of constituents in a DP
(more or less, a Noun Phrase).

 Because of the rhyme with roote

Soote is one of the words for sweet (German süss) (in MoE soot was lost)

Why is it soote shoures (with a pronounced schwa)

 it shows plural

The schwa at the end of soote shows the remnant of the OE strong adjective declension. As you can see we think
that Chaucer inverted the order of soote and shoures to have a rhyme with roote, which is also thus assumed to
have had a schwa at the end (roote). What was the grammatical function of this schwa?
  it showed DAT

We think this schwa may still have been there for grammatical reasons: it shows the continuation of OE DAT
case in masc and neut nouns (-e). Chaucer, however, very often forgets about this DAT suffix and sometimes the
suffix is used where it could have ben present in historical terms. This shows that its original function was lost
and in the confusion free variation (with zero suffix) ensued, after which it was finally lost (all ME schwas are
lost). For those who want to see the suffix, in German there are nice (fossilized) examples:

zu Hause 'at home'

auf dem Wege 'on the way of...'

Dem Deutschen Volke 'to the German people' (inscription on the houses of parliament/Bundestag in Berlin)

This suffix is no longer productive. The same may have been true for Chaucer's times.

Read the first two lines again. Now look at what comes next.

And bathéd every veyne in swich licóur

Of which vertú engendréd is the flour;

Remember: the stress marks (usually!) show pronounced schwas (or unusual stresses)! Why don't older editions
show stress? Because it is not difficult to find using the underlying iambic pentameter pattern. Now let's convert
them to show the place of stress.

And báthed évery véyne in swích licóur

Of whích vertú engéndred is the flóur;

Before we translate it, let's look at phonology: there are too many vowels. évery must be scanned as a bi-syllabic
word. Why was the vowel lost?

 syncope

Syncope is a magical means of achieving the strict alternation of stressed and unstressed syllables. Syncopation
is the loss of a vowel between two consonants if there is a singleton consonant after the vowel: CVCVCV >
CVCCV (every > ev'ry). Syncopation is still active in English (the spelling does not show it, however):

history > hist'ry

hetero >het'ro

family > fam'ly

Note that there is an extra schwa (or at least there could have been one) at the end of veyne 'vein' (from
Latin vena though OFr, the schwa was in all likelihood still there in the OFr word). So how does veyne end up
with one vowel (= how is the schwa lost)?

In addition to syncope there is another 'policing' force in a Chaucerian line: elision, which means that is a
sequence of two vowels the fist one is lost (even across word boundaries): ..CV# #V... > ...C# #V... (veyne is >
veyn' is). So whenever you want to get rid of a vowel try applying elision or syncope. But be careful, like the two
proverbial hedgehogs...

Translation

and (has) bathed every vein is such juice

of which power the flower is begotten (= made).

Note Chaucer's use of virtue 'power'. The word has an interesting history, originally the word meant 'power of a
man, physical strength' (the Latin stem behind the word is vir 'man'). Later the meaning lost the physical aspect
to it and now it means 'moral (if somewhat exaggerated) strength'. Words like virus, viral all come from the
same stem. The English cognate of the Latin words is... wait for it... werewolf 'wolf man' or 'man wolf'. One of
the OE words for man was wer. There is another word having this stem: world (OE werald 'age that a
man/human can live to' (wer + eald 'old').

So what are these four lines about (as is the whole 18 line long sentence)?

 sex

It's sex without our two hedgehogs, of course couched in very subtle Chaucerian terms. We are nearing the end
of March (the dry period/drought; don't confuse it with draft!), and everything is waiting to receive some rain (I
hope I don't have to be more explicit on this) for regeneration.

Note the semantic change in liquor (today 'alcoholic drink', in ME 'any liquid'). This is specialization of meaning
(read the chapter on this).

Bathed is...

 past participle
 non-finite (has no person number

It is past pple because it is coordinated with hath perced. Look at hath. What is this?

 It shows the continuation of OE 3 Sg Pres Indic suffix -th

 It is finite
 It is an auxiliary
Yes, the suffix was later replaced with -s (from the North, read about this in the book and the quiz on OE).
Chaucer knew about this suffix, of course. There is something very interesting we can spot about it. When
Chaucer does use it, he puts it into the mouths of...
 the poor from the North

Yes, Chaucer uses this is a real dialectal marker: speakers form the North (it is not clear where the dialectal
boundaries were) have -s, those from the South have -th. Of course, as always the negative connotations
permeate this distinction: speakers from the North are though of as being poor, uncultured, etc. If only Chaucer
had known what was to happen to the southern suffix -th...

Let's push on.

Whan Zephirus eek with his sweté breeth 5

Inspiréd hath in every holt and heeth

The tendré croppés;

Let's mark the stresses:

Whan Zéphirús éek wíth his swéte bréeth 5

Inspíred háth in évery hólt and héeth

The téndre cróppes;

Let's analyse this phonologically.

Why is there a schwa at the end of sweete? This...

  shows a definite noun phrase
 is the continuation of the OE weak adjectival declension

Note the schwas: they appear in the plural (goode men, the goode men, his soote shoures), and the singular after
a definite determiner (the/his goode man). After ME all traces of this distinction are lost. But, if you wish to see
how this works go to German: ein guter Mann vs der gute Mann.

Note that ME has something that OE lacked: indefinite article (a(n) did not exist), so OE gōd mann = ME a
good man. OE had no definite article either, the demonstrative pronoun developed into the def article, so today
you have a grammatical difference between the good man and that good man. OE was similar to many IE
languages in this respect: the Slavonic languages still lack an indefinite/definite article. Which was did the
indefinite article develop from? Yes, it's the numeral one: OE ān mann > a man, ān æpel > an apple.

Note the tendre croppes. Now you know the schwa in tendre shows plural (and definiteness).

Look at the number of syllables: every is two syllables long because of syncope.

Let's translate:

When Zephyrus also with his sweet breath

has breathed life on every forest and field

into the tender shoots.

(Zephirus is spelt Zephyrus today)

The syntax is a bit mangled. Rearrange the constituents into S V O A.

When Zephyrus has also breathed life into the tender shoots in every forest and field with his sweet breath

This is not the most poetic of translations, but that's not the point.

Note the words: Zephyrus 'The Ancient Greek god of the west wind, supposed to be the gentlest of the winds and
the messenger of spring' (from where spring comes and the Sun rises), eek 'also' (German auch, the word was
lost, but is preserved in nickname < an eek name 'an additional name' where the 'n' of the indefinite article was
reanalyzed as part of ickname), inspire (the original meaning is 'breathe into' from Latin in 'into'
and spirare 'breathe'), holt 'forest, woods' (German Holz 'wood (material)', Russian колода 'piece of thick wood',
do the calculations) (Note: holt today means a place where otters live, made from wood), heath (the same
meaning today: a field where there are no crops; also the name of a
plant (heath), Hungarian hanga, Latin erica, all used as personal names, cf. English Heather), tender 'young',
crop (for Chaucer this was 'shoot, sapling, young plant', not 'harvested plants', as it is for us).

Let's move on.

.....and the yongé sonné

Hath in the Ram his halfé cours yronné,

Now mark the stresses before your mental eyes:

...... ánd the yónge sónne

Hath ín the Rám his hálfe cóurs yrónne,

The translation runs along these lines:

... and the young sun

has run/made half its course in the Aries.

Why young sun? It's the spring sun which is not as strong as the summer sun.
Who is the Ram? It is the OE equivalent of the Aries (the sign of the zodiac under which spring arrives). Why
did Chaucer use Ram? Some would say because Aries was too long a word (it was already in use in his time, of
course). There is a wonderful piece of information that supports this: next to line 8 (check the original
manuscript) there is a Latin remark: sol in ariete 'sun inAries'. This suggests that Ram was probably difficult to
interpret (perhaps the buyer of the manuscript was French and would have needed this remark to understand the
text).

Let's see the morphology. Why is there a schwa in the yonge sunne and his halfe course?

 yong and half is in a definite noun phrase

 yong and half appear in its weak form

What is yronne?
 pst ppl
What is this y- before ronne?
 The remnant of the OE prefix ge- (which sometimes appeared in pt pple)

Which is the only word in English which preserves this ME y- (OE ge-)? Enough, the pt pple of a now non-
existent OE verb nugan 'to suffice'. This is a real relic. If you speak German, you know that cognate
is genug (but it also exists in other forms: genügend 'sufficient').

Why is there a schwa at the end of the noun sonne 'sun'?

 this is part of the noun

Yes, it is part of the noun. The OE predecessor was sunne. This was a weak feminine noun. It seems that ME
also had long (identical) consonants. We can't be sure what the late ME (Chaucerian) pronunciation was, but we
have independent proof that early ME did have these....

 geminates

Now if the sun has completed half its journey in Aries, which day of which month does Chaucer have in mind?

 1 April

Fascinating. April (at least in the past) used to be the first rainy month after a long dry spell ending with March
(the drought of March).

We haven't yet reached the main clause (this is still the subordinate one beginning with whan).

Go back to the very first two words in the first line: whan that. What is wrong with this in MoE? Yes, no
sentence can start with these two words. For those who are syntax buffs this will be known as the 'doubly filled
complement filter' (you can't fill with overt words both the specifier of the CP and its head). Disregard this if no
familiar bells are rung, but consider this:

(i) Peter is the man who I look up to

(ii) Peter is the man that I look up to

(iii) Peter is the man 0 0 I look up to

(iv) *Peter is the man who that I look up to

(i) to (iv) show sentences with a relative clause (subordinated to a noun), a structure very similar to a subordinate
clause (subordinated to a verb). Observe that the relative pronoun (here: who) can never be followed
by that. Leaving both out is grammatical, but having both of them as overt words is disallowed. The exact
conditioning is not fully understood, but it seems ME had no such constraint.

We still have a few lines to cover.

And smalé fowlés makén melodyé,

That slepén all the nyght with open eyé, —

Show the stresses.

And smále fówles máken mélodýe,

That slépen áll the nýght with ópen éye, —

I'm sure you can see the morphological questions coming your way.

Why is there a schwa in smale?

 The adjective is strong

 It shows plural

 The noun phrase is indefinite

But let's see the translation.

and small birds make melody (= sing)

that sleep with open eyes all night

Why is there no -s at the end of eye? The OE noun was ēage and one of the very few weak neuter nouns. Yes,
there used to be a n here (eyen, and Chaucer does use it on occasions). However, word final inflectional -n was
slowly being lost in ME, so the word in the plural could be found as both eye [ɪjə] and eyen[ɪjən]. Actually the
Sg of eye was also eye. (for those who speak German: Auge is still a weak noun whose Pl is Augen).

Now comes a difficult one? Why is there no -e (schwa) at the end of open in open eye?

 It was deleted by elision (opene eye)

What is slepen?

 3 Pl Pres Indic
Which word is ME fowl continued as with a difference in meaning?
 Fowl (now referring to farm birds)

The word underwent semantic specialization (the ME word referred to any bird, cf. the German
cognate Vogel which still means any bird).

There is a structure which can be contributed to Romance influence: maken melodye (a long phrase for 'sing'
probably modelled on French 'faire une melodie'). Note that ME can also provide interesting information of Old
French: melodye rhymes with eye, so we must suppose that there was a schwa at the end of the OFr word. It
comes from Latin melodia, which shows that Latin unstressed -a went to to schwa in OFr and was ultimately
lost). ME is the grave yard of dead OFr words. Actually even the ME schwa in melodye was lost in the
transition between ME and eMoE.

And now a side remark:

So prikéth hem Natúre in hir corágés —;

Let's show the stresses.

So príketh hém Natúre in hír coráges —;

Let's translate.

Nature excites them so in their hearts in this manner

Priketh is of course 3 Sg Pres Indic. Note that here the verb means excite, arouse, not stab. Try analyzing
why prick today means what it does (cock, penis -- excise my English).

Hem is the continuation of OE hem 'them'. Where does them comes from? From the North, it's a Scandinavian
word ultimately, which had not yet reached the South by Chaucer's times. Where does hem survive in MoE? It
survives as 'em after verbs:
Kill 'em

Don't touch 'em

Take 'em away

Hir is the ME (and OE) word for their. Yes, their is another Scandinavian word. Chaucer does not use it for any
of his southern pilgrims.

Courages is French word meaning 'hearts'. Today courage can only be used metaphorically, for mental strength,
but not the organ. Why did Chaucer use courages? He needed a word to have a rhyme.

I hope you can appreciate what a fundamental change happened in the 300 years that separate Chaucer from OE:
even if we disregard the phonological and morphological changes, English is now thoroughly imbued with Old
French and Latin words. English is now starting to look as we know it. The vocabulary (= lexicon) had lost a
tremendous amount of OE words (some of them were 'ousted' by Old French, some were just forgotten as it
usually happens in the transition of language between generations). All in all, English now sounds less like
German, although its is still a Germanic language.

And now we finally enter the main clause.

Than longen folk to goon on pilgrimages,

And palmeres for to seken straunge strondes,

Let's find the stresses and underline the schwas.

Than lóngen fólk to góon on pílgrimáges,

And pálmeres fór to séken stráunge stróndes,

Look at the translation.

Then people long to go on pilgrimages

and palm leaf carriers to look for foreign countries

(Note: palmers are those who had already been to the Holy Land and come back with a souvenir, a palm leaf.)

Let's analyze it morphologically.

Longen is...

 finite (3 Pl Indic Pres)

 the V of the S (= folk)

Why does folk have no plural -s? (Not even in MoE)
 It belonged to a class where the Sg looked the same as the Pl

Indeed, folk is like sheep, deer, and so on (but not like information, work, advice which have no plural and occur
only in the singular). This is not a phonological matter, these nouns belong to an OE nominal class in which the
Nom Pl looked the same as the Nom Sg (the reason behind this may have been phonological, but that's buried in
history).

Why is the plural of palmeres [palmərs] (not [palmərəs] or [palmrəs] via syncope)? Compare it
to strondes [strɔndəs] (not [strɔnds]). This is a phonological question.

 The plural -es had no vowel in words longer than a mono-syllable

It seems that palmeres was just too long phonologically. If you speak German you may now that the genitive
suffix -(e)s is [əs] after mono-syllables, but [s] after words that happen to be longer: Stadt 'city' -- Stadtes 'of the
city', Wagen 'carriage' -- Wagens 'of the carriage' (not Wagenes).
What is the schwa doing at the end of straunge? It...

 is plural
 is a strong adjective

Look at the semantics of straunger 'strange'. This is an OFr word, of course, but it's meaning has shifted from
'foreign' to 'odd' (some may call it pejoration). You can no longer refer to a foreigner as a stranger. You can, but
then a jocular twist is implied as in: Hello, stranger! Note that in French étranger still means 'foreigner' (What
happened to the s in French? Of course, it was lost (we even discussed it in class), but it didn't turn into é. If you
are interested write to me. We must uphold an air of mystique here)

Look at strond (it could also be found as strand). The original meaning was coast, beach, hence land, hence
country (for Chaucer). The word in MoE is almost lost, bu tit survives in the phrase leave someone stranded =
leave someone in the lurch/high and dry/deserted/cheated on, etc. Can you see why? If you end up stranded, you
ship suffered shipwreck, and you were lucky enough you ended up on a beach, and then suffered very likely
death unless saved, which didn't always happen. The word is well-known in Hungarian as strand [ʃtrɑnd]
meaning 'beach'. This word doe snot come from English, but from German (which has this word as a cognate
with English strond). Note that English has this word in street names, as you have in London in the name The
Strand. Why is this called The Strand (Links to an external site.) if it is not anywhere close to a river bank? Well,
the Thames has meandered away from it in the last 500 years. It's not only language that changes.

What are goon and seken?

 non-finite verb forms

 infinitives

We have seen that the infinitive suffix was -en [ən] and it also alternated with [ə] and sometimes was totally lost
(seken ~ seke ~ seek). What is the name of the phonological rule that deletes the schwa is goon [gɔːn], not
[gɔːən] (the same we saw in palmeres being palmers. Note also thanne 'then' which scans here as a mono-
syllabic word.

 apocope

This is apocope (the loss of C's or V's from the end of a word that cannot be explained with other phonological
means).

There is a syntactic feature that no longer exits in MoE, the for to infinitive. It was lost (syntacticians probably
have an answer to this and probably has something to do with case assignment, which we cannot discuss here). If
you speak German you know that German still has it in the form of um zu 'to, in order to'.

Let's move on.

...To ferne halwes, kowthe in sondry londes;

Let's mark the stresses and the schwas.

...To férne hálwes, kówthe in sóndry lóndes;

Let's look at the translation.

(they still long to go on pilgrimages) to far-off shrines (that are) known in various countries.

We have quite a lot going on here.

ME ferne > MoE far (German fern) The phonological story that turned ɛ to a (later to ɑ) is interesting. Look at
some other words:

ster ɛ > star ɑ

ferm ɛ > farm ɑ

fern ɛ > far ɑ

mervel ɛ > marvel ɑ

compare these to:

ferry ɛ

sherry ɛ

terrier ɛ

perish ɛ

(with no changes)

The change happened after ME. The vowel ɛ was lowered by r to ɑ when r was

 followed by a C or if it was word-final (= r was in coda)

This rules is known as farm-rule. Now you know why. The word farm comes from ME ferm, which is an OFr
word, which in turn comes from Latin firma 'steadfast, dependable, established'.

Do you remember the OE version of Lord' Prayer? We had the pt pple halgode 'sanctified, hallowed'. Well, this
is the same verb. OE <g> is found as [w] before bak vowels in ME (spelled <w>). The phonology behind this is
interesting but we can't go there. This comes from OE halgas 'shrines' > ME halwes. Why wasn't the schwa lost
in the plural suffix?

 The stem is mono-syllabic

Have you ever wondered what lies behind Halloween? Yes, it is this OE stem. E'en is even (an alternative form
of evening). It translates as 'hallowed evening'. If you are asking why the [v] disappeared, I will have to urge you
to take a specialization course in historical phonology.

Now we arrive to kowthe 'known'. One of the possible meaning of the verb know is 'be able to, know how to'. In
many languages the word that developed into the modal can actually originates in the verb know. Yes, English is
one of these: the modal can and know can be derived from the same Germanic verb. Even the C's are the same
(the <k> in know was deleted because of a phonological constraint not very long ago). The verb know is a strong
verb (you now know why): the past is NOT formed by adding -t-/-d-. The verb can, however, is a weak verb: the
past is formed with -d (could) (Note: the -l- is unhistorical, it was added by analogy with will, which doe shave
an -l- in the past). The pt pple (and passive participle) is also expected to be formed with -t-/-d-, but in a special
subclass of verbs it can also be th. So far so good (now we know where the k and th come from in kouthe), but
where did -n- go? This is a phonological issue: -n- was deleted before voiceless fricatives in pre-OE (so a very
long time ago). And á presto we have kowthe. So modal verbs did have a participle at one time. German modal
verbs still have a past participle (and infinitive as well): können 'to can', gekonnt 'kowth'. The story is long and
interesting, this is just the tip of the iceberg.

This verb has a long provenance. Calculate from english what Latin and Greek will have? Yes, they must
contain g n (cf. Latin gnosco 'I know', agnosticus 'not knowing', etc). The Slavonic languages have a z from g,
hence znat' 'to know'. This verb is as old as IE.

I hope I still have you attention. Modals in MoE have no non-finite forms, so they are always finite (= having
person number tense mood). This wasn't always the case (see kowthe for the pt pple). Did the verb can also have
a present (-ing) participle? Yes. But it crossed over to the class of adjectives (actually ing-participles can also be
used as adjectives): cunning 'knowing what to do'.

Back to Chaucer. Why does Chaucer write an -e in kowthe (but none is sounded)? Note that the pt participle is
just like any odd adjective.

 It can be analyzed as the plural of the adjective (the schwas is elided)

Sondry is spelled sundry today meaning 'various' (German sonder 'special').

Londes is lands (for phonology buffs in some dialects of English the <a> followed by a nasal was raised to <o>,
hence lond, mon, sond, strong; only strong survives from this dialect). Why is there a schwa is londes? Sorry,I
know you know (but just in case you don't remember: the noun is mono-syllabic).
We are nearing the end of the first sentence of The Prologue.

And specially, from every shires ende 15

Of Engelond, to Caunterbury they wende,

Now let's mark the stresses and the schwas.

And spéciallý, from évery shíres énde 15

Of Éngelónd, to Cáunterburý they wénde,

Let's translate.

and especially from the corner of every shire

of England to Canterbury they go

Don't think that Éngelónd was stressed on the last syllable. The stress was on the first syllable as it still is, but the
word was still considered a compound word consisting of two words (similarly to MoE bláckboard, líghthouse,
Whíte house, etc.). The second part still had some prominence. Now observe Cáunterburý and spéciallý. It
seems that ME preserved a long vowel for OE word final [i:] (or [ij]), spelled <y> in ME. This long ME [ij]
survived into eMoE as well (so it wasn't lost as the schwa was). This is also a long story.

Why is there no schwa in Caunterbury?

 It was deleted through syncope

Why is there no syncope in Engelond? It seems that syncope does not work across words (#Enge##lond#).
Ultimately the schwa was lost but possibly after this became a mono-morphemic word (#England#).

You know what is going on in every.

Loot at shires. It must be the genitive shire's. The apostrophe was not going to be used for another 400 years,
so shires could be GEN Sg (shire's), NOM Pl (shires) and even GEN Pl (shires'). There is an interesting
morphological peculiarity that spans a number of thousand of years. The only OE case suffixes to survive into
MoE are the GEN Sg -s and the plural -s. Note, however, that these suffixes were confined to specific classes
(strong masc and neut nouns), from where they were generalized to the rest of the classes. Feminine nouns never
had this -s, and neither did the weak nouns (of any gender).

A little curiosity that you know about English but have never known quite how to treat it. Consider:

a ten mile journey (never a ten miles journey)

a ten foot ledger (never a ten feet ledger)

a five hour flight (never a five hours flight)

What is going on here? If plural nouns premodify another noun they can't have the plural, in any form). This is
actually an invisible continuation of the OE plural genitive suffix -a (e.g., OE daga 'of days', banena 'of
murderers', heortena 'of hearts', etc.). This -a surfaces as -e (a schwa) in ME. What happened to every word-final
schwa after ME?

 it was lost

All is fine, language works as it should. This explains this modern day oddity.

Let's march on. What is wende?

 3 Pl Indic Pres

Oh yes, for Chaucer wenden existed in its infinitive (wenden), the present tense forms (I wend(e), thou wendest,
he wendeth, we/ye/they wend(e)), and also in the past it had both regular and irregular weak forms (I
went/wended, thou wentest/wendest, he wended/went, we/ye/they wend(en)). What survived is the past (the
present having been lost) which joined in with go that lost its past, so now we have two halves of a defective
paradigm joining forces: go -- went --gone. This phenomenon is known as suppletion (a word supplying its
missing inflectional forms with a word that is not its cognate). For those who speak German, the cognate verb
is wenden 'turn to, go' which has all its forms.

You can ask where the weak -d- is in the past of wenden in ME. It's there, but was not spelled and was probably
impossible to tell apart from the -d of the stem: I/he wended, thou wendedest > wenddest > wendest. The
devoicing of -d word finally is another interesting question, giving us MoE went. We can't go here now.

You can see that ende rhymes with wende. Why is there a schwa at the end of ende?

 the schwa comes from OE

Remember: only adjectives can actively hover between weak and strong forms. Nouns are what they are, the
strong/weak declension is a historical difference (and a matter of analysis).

Notice they (also spelled thay). The interesting thing is that they had its Acc form (hem), and its possessive
form (hir). The OE for they was hie. How come this is already they for Chaucer? Well, not all guest from the
north arrive at the same time. This is also a Scandinavian word, but it arrived a few centuries earlier
than them and their. Of course, hie for they is also found in ME.

Let's move on (we could spend a lifetime discussing this).

The last two lines of the first sentence read as follows:

The hooly blisful martir for to seeke,

That hem hath holpen whan that they were seke.

Let's show stresses and the schwas.

The hóoly blísful mártir fór to séeke,

That hém hath hólpen whán that théy were séke.

Let's translate.

... to seek the holy blessed martyr

that has helped them when they were sick.

We all know who they are keen to visit. We also know that for to is no longer used in MoE. We also know
that seeke is infinitive and can have a number of alternating forms: seeken ~ seeke ~ seke ~ seek. That (a relative
pronoun) was already used as it is in MoE (this would have been impossible in OE). You also know that whan
that is also impossible in MoE (we talked about this above: that can no longer be preceded by a +wh word
like when, etc).

Now look at holpen. Yes, it does look like a strong verb (it has no -d/-t in the participle). But what happened to
it? It crossed over to the weak side, so now we have it as help -- helped -- helped. If it had survived we would
(probably) have it as help -- halp -- holp. Note that in German the verb is still strong: helf- -- half -- (ge)holf-.

The past tense of be is were. Why is there no schwa here? It should have one as it comes from OE wǣron.
  Schwa was lost through apocope
 Chaucer chose a form that could have a mono-syllabic pronunciation

Direct inheritance vs. borrowing

We have seen that the relationship between any two languages can be approached from a number of
perspectives. Language A and language B can be studied from a number of perspectives.

They can be studied as a typological phenomenon: how the two languages choose to assemble morphemes (=
packages of information) into larger strings.

There are isolating languages like Chinese and English that by and large have free morphemes carrying
packages of information (e.g., I have never been actually seen eating sausages in my
life where have and been carry perfect aspect, but are free morphemes. You can insert words between them, as
we have just done above).

Some languages are agglutinative and like assembling morphological information into long, superwords, like
Hungarian and Turkish (recall the example we looked at in class from Hungarian: megetetgethették 'they may
have been making someone feed it on and off', analysed as meg|e|tet|get|het|t|e|k with the following morphemes:
meg- 'perfective', e- the lexical stem for eat, tet-'causative' (make so do something), het- 'modal similar to
En can', t- 'past tense', e- '3 Sg = he/she', -k 'plural' (= he/she + = they), which also lengthens the vowel, hence -
é-). A wonderfully complicated machinery! Note that here the stem -e- 'eat' is actually not a free morpheme.
Hungarian is usually said to have free stems after all the suffixes and prefixes have been cut away (in our
example -e- is not a free lexical stem). Note the word usually though.

Some languages are inflectional (or inflecting) like Latin or the modern Slavic languages may have rather long
words with suffixes and prefixes and typically a bound stem, meaning that it can not be found in isolation
(similarly to Hungarian above). The Latin stem for earth is terr- (note the hyphen after it showing it to be a
bound stem), always appearing with suffixes: terr-a, terr-a-m, terr-ae, terrarum, etc. The lexical stem for earth
is never found as terr.

But this looks like Hungarian, right? Well, no, because Latin (like all Indo-European languages that sill work
like it) have fused morphemes (older books refer to them as portmanteau morphs -- if you speak French, you
will know that portmanteau means coat hanger). What's this? Look at terra. The suffix -a is actually standing
for feminine, singular, nominative. The suffix -a is like a coat hanger onto which you throw a number of items of
clothing and at some point you will not know where one ends and the other begins. Look at terram: you can say
-a- shows that the noun is feminine (this is usually claimed for this suffix on the basis of the other declensional
classes of Latin), -m shows that it is accusative singular. And so on. Compare this to Hungarian. Hungarian
usually has suffixes that are not fused suffixes: the -t- above shows past and that's it, -k shows plural and that's it,
-tet- shows causative, etc. Note again the word usually... Hungarian also has fused morphemes, rest assured.
We're not going there.

We are talking about typical features here. This gives us the three broad classes of languages: isolating,
agglutinative, inflecting. Agglutinative and inflecting languages are known as synthetic (as they synthesize
material into long words). Isolating languages are also referred to as analytical (for obvious reasons). There is
one more type that we will not discuss here: polysynthetic (Inuit and Australian aboriginal languages are famous
examples). This is much more complicated and varied than presented here, of course.

Have you noticed that we have not yet said a single word about whether such languages are genetically related or
not. Why? Because typological linguistics is not interested in this. So what is?

Genealogical linguistics looks at whether similarities among words (morphemes) is accidental or systematic
irrespective of how these are arranged in a word. Note that we have compared English to Chinese, Hungarian to
Latin, and so on, and they have nothing to do with each other genetically. This is just fine. Typological
linguistics is not interested in the 'genes'. Before we continue, try answering the following question based on
what you have read so far in the compulsory book.

What do we mean by genetic/genealogical relationship among languages? That...

 they are descended from a common 'mother' via sound changes

 they can be related to each other with the help of common words going back to the 'mother'
Very good. Biological genes have nothing to do with genealogical linguistics, and rightly so as this aspect of
linguistics examines the languages not the people who speak them. Can you see that both typological and
genealogical linguistics have nothing to say about us, speakers of a language. Beautiful. We are actually looking
at an abstract web of relationships across morphemes in different languages. Languages are abstract means of
communication, a system that is that can be studies independently of those who use it. If you find this
discomforting look at physics. Physics describes the workings of the universe with no recourse to us, humans.
Actually it is said that the universe does not need us to function. Whether we agree with this or not is a
philosophical question.

I hope you can see why we can study dead languages like ancient Egyptian. Exactly! We don't need its speakers.

Which branch of linguistics can shed light on why (and to what degree) languages are related?

 phonology and morphology

Actually phonology rules supreme, but it can't do without morphology and semantics.

Morphology is needed not because how morphemes are structured but because related languages will show
similar usage of similar morphemes (Latin -ae above is the dative in Latin, and the same historical suffix was
found in Slavonic and Greek and Germanic at a certain stage for the same function, but not Hungarian. You may
ask where this suffix is in English. Well, it was lost, but Old English still had a reflex of it in certain declensional
classes).

Why is semantics needed? Because comparison among related words (or words thought to be related) must
assume that the meaning that we can find today must contain a common core or a development of it. Semantics
is a capricious thing, but necessary when we establish compare morphemes in the wish of establishing a
common/mother morpheme.

Let's look at some examples. English has head, for which German has Kopf. If you have read the suggested
material so far, you will say the cannot be related to each other because no sound rules can establish such a
relationship. True! Perhaps you are looking in the wrong direction. Given the set of correspondences between
English and German, German k should correspond to k in English, and pf to p. So what is this word? It is cup.
How? What can the common semantic core be? Possibly the notion of a container (which contains your brains or
some tea). If you speak French, you know that the word for head is tête, which goes back to Latin testa meaning
bowl. Wonderful! You can see the same principle here. But what happened to Latin caput 'head'. Nothing has to
survive, you must understand this. Caput does survive actually as French chapitre meaning little head(ing) in a
book, i.e., a chapter (indeed, the word chapter is an Old French word).

When you read the book you will see that we always must consult semantics when we try to devise a mother
language.

Up to now we have seen that languages can be looked at from the perspective of typological linguistics
and genealogical linguistics.

Is there anything else? Yes. It is areal linguistics which looks at how languages can influence each other in a
common space, especially by having close connections among its speakers through conquest, commerce,
interest, and a combination of these. If you think about this, this is rather self-explanatory. Hungarian has not
been influenced in any way by Japanese. Why? It's just too far away. However, Hungarian has been heavily
influenced by the IE languages around it, especially German and the Slavonic languages, so much so that in the
19th century the scientific discourse at the time went on about whether Hungarian was a quaint dialect of
German or one of the Slavonic languages. Over 60% of Hungarian words are of Germanic and Slavonic origin
(with some hints of Turkish and a number of other languages). What do you expect of a language when it has no
native words for Tuesday, Wednesday, Thursday, Friday and Saturday? All these are of Slavonic (and ultimately
IE) origin.

So why is Hungarian not an IE language?

(Careful: not because it is agglutinative (because languages can become agglutinative over time.))

 When you remove all that you think is IE you are left with words that have no IE origin
 You can explain why IE words entered Hungarian

 You cannot explain why certain words have always been part of the language
By and large that's it. After you have removed the layer of supposed IE words, all that remains must be
something that you cannot explain with borrowing. This is Finno-Ugric. You will say that all the basic words
(numbers, parts of the body, etc.). Yes and no. Even basic words can be lost and substituted with non-native
words. A famous example is the vulgar Hungarian word for copulate, have sex (shag for those who do not blush
at this word). The stem is basz-. You can't possibly say this is a new word, after all people have always had sex
and have never blushed at the prospect of talking about it. This word, however, is a Turkic word (= from one of
the Turkic languages), the original meaning was hit, push. The Hungarian expression Ott baszták a rezet egész
nap (meaning: They have been idling away time there all day, literally: They have been hitting brass all day
long). This expression preserves the original meaning. Of course, you can see that having sex and hitting is not
really that far away from each other semantically. Recall the English word fuck which can also mean 'hit' as in
the expression You wouldn't see it, not even if it came up to you and fucked you in the face. Recall that fuck has
its Latin cognates pu(n )g- meaning 'hit' (this Latin word was borrowed from French or Latin itself in the
Renaissance as pungent, meaning strikingly strong, piercing).

You may ask is it possible for the speakers of a language to completely forget their 'original' native language and
adopt a variety of a neighbouring language (for example by means of being subjugated)? Yes, and this has
happened to a number of languages. This is known as language switch (you switch to a more dominant,
prestigious language and with time your progeny will no longer speak the original language) Which one of these
languages show such a switch?

 Irish English
 Bulgarian

Very interesting, isn't it? Irish English is a variety of (English) English having been transported over to Ireland
and superimposed on an originally Celtic language. Note that Irish Celtic is now being revived (there are still
speakers, very old admittedly, who still speak it). The same happened to Welsh English. As a matter of fact a
variety of Southern English was superimposed on another dialect of Insular Celtic (spoken in Britain). Note that
the native Celtic of Wales is now also being revived.

What about Bulgarian, a Southern Slavonic language? The speakers originally spoke a Turkic language but over
time adopted a Slavonic language. The word 'Bulgarian' actually shows the origin of these speakers through a
well known alternation: Bulg- ~ Volg-. Yes, these people originally come from around the river Volga in today's
Russia. Is this Turkic language being revived in Bulgaria? No. No speaker can be found who has had any
appreciable bit of memory of this language. The language was forgotten over a thousand years ago. But note that
languages do not die because the linguistic system dies, they die because their speakers die (or choose not to
hand them over to the next generation).

What about Hungarian? Two options present themselves (no one contests it not being an IE language!): (i) it
may have been originally a Finno-Ugric language that whose speakers switched to a Turkic language (with a lot
of words surviving from the original Finno-Ugric words) to which later IE words were added, or (ii) Hungarian
was originally a Turkic language whose speakers switched to a Finno-Ugric language at some point in the
history (to which later IE words were added). The consensus among language historians is that Hungarian is a
Finno-Ugric language that adopted a substantial amount of words from a number of Turkic languages (before it
came into contact with IE languages). Why would a language historian say this? Because of reconstructive
phonology: the sound regularities show that it is more plausible to compare Hungarian to Finnish, Estonian and
the rest of the Finno-Ugric languages than it is to compare it to a Turkic language.

English offers a wonderful case study. If you come to think about it English and French are very similar (even
the spelling is identical for very many words, like the word coin, but spelling is not essentially part of linguistic
comparison). It doesn't mean of course that if you speak French it is easy peasy lemon squeezy learning English,
but then if you speak German you will never learn Russian only by observing German (despite both of these
being IE languages). A major portion of English vocabulary is of French origin, but what do we mean by this?
You know by now. The French layer of English can be explained with areal linguistics: French was
superimposed onto English after the Norman Conquest in 1066, it has always been a language of finesse,
refinement, culture, learning, perfumes, you name it.

After you remove the French layer of the vocabulary by analyzing it away as borrowing through
superimposition of a dominant language, what stays behind? Well, the Germanic part of the vocabulary, the part
that you cannot explain by saying that it must have been superimposed onto English at any stage. Why not?
Because English is a Germanic language and nobody can deny that. Why? Because historical phonology says
that English is more plausible to compare to the Germanic languages than it is to compare to French. By now
you know that English through Germanic is related to French (via Latin), but that is a very distant relationship.
In other words, the relationship between English and French is very shallow, as it were: the relationship between
French words and English words of French origin is just skin deep, and can be explained as borrowing. If you
read the chapter in the obligatory book you know what reconstructive phonology can do and how far it can reach
into the history of languages.

We will see that English is very interesting though from the perspective French. Why? Because it preserves
some features from Old French that have been lost in Modern Frech. For example, French has no dʒ or tʃ today
in words like journal or chanter (it has ʒ and ʃ instead). How do we know what Old French had? From Middle
English (through Early Modern English to Modern English) journal and chant. You can say that English is the
graveyard of dead French words. Note that if Modern English borrows words from Modern French, it never the
Old French sounds dʒ or tʃ, cf. chef, sachet, genre. It has the Modern Frech sounds, of course.

You can see now that a similar metaphor can be said for Hungarian: it is the graveyard of dead Slavonic,
Germanic and Turkish words. If you speak a Turkic language you know that kapu 'gate' in Hungarian is loan
from Turkish. In Turkish the words is kapı (the letter ı shows a back high unrounded vowel similar to Polish y or
Russian ы. You may ask which language shows the original pronunciation? Well, it is Hungarian. The Old
Turkish word had u , which later lost its rounding.

Look at Budapest. The second part of the words comes from Slavonic pešti (from *pekti-, recall: * shows
reconstructed words) meaning fireplace. If you speak any of the Slavonic languages you will recognize the word
as печь, peć, peč, piec, etc. Budapest has always been famous for its warm baths, furnaces where iron was
smelted. For those who speak German, you will appreciate that Pest's name is Ofen 'fireplace' in German (just
look at old maps of the city, and yes, Ofen is the same word as oven in English). Allegedly, Buda is a Celtic
word, meaning.... yes, fireplace. You will appreciate that another city in Hungary preserves the Slavonic word
for fireplace: Pécs petʃ (in the South of Hungary). This word comes from a Southern Slavonic dialect and shows
a phonological development of *pekti- (which goes back to the ultimate stem *pek- 'bake', a stem familiar to
those who speak a Slavonic language).

So, let's introduce some terminology (this may be familiar for those who started reading the book for the exam).
If two genetically related languages share common words then these words can be traced back to an original
(usually reconstructed) word that is the mother word (morpheme),from which the daughter words derive through
regular sound changes. The formulation of regular sound correspondences was a revolutionary achievement of
the 19 century. Before that (insightful) remarks had been made, but never at the level of what comparative
historical phonology achieved in the 19th century.

If you have Latin foris, Greek thyre, Slavonic dveri, English door for (broadly) an opening in a human
edifice (= door) then you can try reconstructing the IE mother word. This should sound familiar to you now: the
IE word is reconstructed as *dhwVr (with V being a vowel that we are not interested in now). IE *dh gives you
Latin f, Greek th (which was an aspirated t, not an interdental/dental fricative θ), Slavonic d, Germanic *d (from
which English inherits door). This set of correspondences can be shown as:

En d ~ L f ~ Gr th ~ Sl d

It is not only these four words that show this correspondence, it is true for every word having this IE sound (of
course, life is more varied than this, there may be special environments, etc, but this need not interest us now). If
other words, we can say this:

IE *dg --> L f, Gr th, Sl d, E d (through Germanic of course)

IE is mother words, it has a number of daughters (direct, or indirect through another mother). This is just line
human genetics: your forebears ten times removed from you gave you a set of genes that you carry. So, in a
sense you are your great-great-grandparents daughter/son through a number of intermediate mothers). (I hope
you have noticed that historical linguistics is gender biased in its description: there is no male input here, but this
is just a tool used in the description).

So, L foris, Gr thyre, Sl dvori, En door are all cognates (from Latin cum- 'together' and the stem gno- 'being
born', so cognates are words that are born together as it were, they are descended from the same mother, etc.).
This is another useful metaphor. Cognates only exist in a mesh of relationships established by regular sound
correspondences and possible meanings that can be established for any such pair. The above example of
'door' is quite self-explanatory. Sometimes the relationships are more difficult to establish. But remember: if you
want to establish a set of cognates you must take into account both regular sound correspondences and meaning.
Bear this in mind as you are reading on.
This is interesting, right? Do we mean to say with this that if a language borrows a word from another language,
it is not a cognate? Yes!

Can you see which word English borrowed from Greek thyre and Latin foris?

 thyroid
 foreign

This is interesting. English has thyroid 'a gland on your throat in the shape of a door/a protecting shield' from
Greek and foreign from Latin (via Old French). But why does thyroid not show any sign of Grimm's Law, etc?

Because it was borrowed by English, not inherited through IE (via Germanic). This word arrives some 3
thousand years too late to be affected by Grimm's Law.

Can you see that now we have two words in English from the same IE word *dhwVr- that arrived through two
different channels into English: (i) one has always been there/it was inherited by English from West Germanic
which inherited it from Common Germanic which in turn inherited it from IE: this is door, (ii) the other
(thyroid) was borrowed from Greek at some very near point in the past. So door and thyroid are...

 doublets

The same is true for foreign from OF (via Latin foraneus literally meaning outside the door).

Please observe again that the relations between door and thyre is much deeper than the relationship between
Gr thyre and En thyroid. What do we mean by 'deep'? Well, the number of rules that is required to
generate door and thyre from IE *dhwVr-. The relationship between En thyroid and Gr thyre is only 'skin deep',
it is too obvious as it were. En thyroid and Gr thyre will not show anything about whether En and Gr are related
genetically. It only shows that have been in contact (this is areal linguistics). You can ask when did English
come into contact with speakers of Ancient Greek? Never. So how can it have borrowed thyroid? Through books
and learned borrowing. Greek has been a fruitful source of loans for all languages in Europe. The name of such
loans or borrowed words is learned borrowing (you simply take a word from a dictionary and start using it).

Recall the Hungarian case of Pest. We said it was borrowed from a Slavonic language (from two sources
actually, once as Pest and once as Pécs). Is this learned borrowing? No. This comes from a spoken source:
Hungarian speakers borrowed a word that was heard in speech (as opposed of thyroid above which comes from
a dictionary). Such a case is known as popular borrowing. What is meant by this? Was the word popular? Well,
it may have been, but the term goes back to popular meaning as used by the people, showing the effects of sound
changes applied to a source word. What's this? We said that the reconstructed Slavonic word for fireplace is
*pekti-. This is reconstruction. The Eastern Slavonic languages have from this *pekti- the word pešti with the
usual change of palatalization of kt to št ʃt before i; later št changed to tʃ), the South Slavonic languages have
from *pekti- through palatalization peć tɕ or peč tʃ). So Hungarian shows what speakers of different dialects of
the same language can do to the same reconstructed word (*pekti). In this manner: Pest and Pécs show the
popular (= as used by speakers) development of the same historical word in Slavonic. So Hungarian has once
again proved to be the graveyard of dead Slavonic words.

You will ask: is it acceptable that palatalization can happen is two different ways? Yes. Languages are masters at
solving problems: kt can be palatalized in a number of ways.

So once more: what do we mean by popular loan?

 A word that shows sound changes associated with changes happening across generations in a donor
language

This is getting ever more interesting. Se we have inherited words, learned loans and popular
loans. Sometimes that same word can enter a number of times. Then we have doublets.

Go back to Hungarian Pest/Pécs. Will you find a learned version of these words? Of course not, we have no
dictionary of Common Slavonic. We do actually, but that's an etymological dictionary used by quaint people,
which certainly does not have the prestige of Ancient Greek. Does Hungarian have an inherited word from
Slavonic? Of course not, because it is not a Slavonic (= IE) language.

Let's go back to English. Analyse chief and chef. Both come from French but at different times. Which one
entered English earlier? Recall our discussion on French tʃ and ʃ.
  chief

Indeed. Even the vowel is very different compared to what we have in chef. So what are chief and chef?

 popular loans

Both are popular. They show the effects of generations of French speakers: chef shows a development of Old
French chief. But then what is OFr chief a development of? Which language do you have to go back to? Yes:
Latin. Very good, but then which Latin word? You know that French has twisted the Latin words to quite a
degree of mastery through sound changes. The Latin word behind this is capitulum 'little head',
from caput 'head'. Try to work out the rules fro yourselves.

So OFr chief shows all the rules associated with OFr inheriting this word from Latin, and Modern
French chef shows all the changes associated with the stages postdating OFr. Wonderful: English has a snapshot
of both stages of French. It has both words as popular loans.

I'm sure you are wondering if English has a learned loan as well? Of course, this is the word capital (as the main
city in a country, and as money needed for investment), Capitol (Hill), etc. These words were simply taken over
from a dictionary of Latin by erudite speakers in the past.

So we have doublets from different sources: capital, chief, chef. None of these is inherited of course.

Now take a leap of imagination :) If Latin has caput what would you expect in English? I hope I can hear your
brains whirring away now at this. It's not easy. Do the calculations!

 head

This is not easy, but you certainly are looking for a word beginning with h. The rest is more difficult, but OE
helps: heafod 'head' pronounced hæɑvod. If Latin has p, the Germanic languages should have f (which is voiced
intervocalically). In some dialects of English, this intervocalic v was lost after OE giving us head.

(For the hard core specialists: if Latin has t in caput, you would not expect d in the Germanic languages, you
would want to have θ and its derivatives in later stages. Let me just say one word: Verner's Law. Try now.)

So we have: head, capital, chief, chef. Pick off the correct answers.

 head is a directly inherited word from IE

 chef is a popular loan from Modern French
 head, chief and chef are doublets

 the relationship between chef and chief is daughter to mother in French

 head (via Germanic) and caput in Latin are cognates

Lets nitpick now. The Middle English word chief (from OFr) pronounced tʃ(j)eːf gives MoEng chief pronounced
tʃɪjf. What is the relationship between these two words in English?
 mother to daughter