The moving red eye

Emotions are not universal. They are expressed in cultural scripts reflected in language, sometimes over wide areas, but drawing on very diverse sources. The metaphors used to describe feelings often can focus on particular body parts as the locus of an emotion. In the Kwa language Ewe of Ghana, some emotional expression – those roughly corresponding to jealousy, envy and covetousness – are built on the image of a red/and or moving eye. Here some examples:

Examples for emotions expressed with ‘eye’ in Ewe (Ameka 2002: 28)

What is special about these expressions is that they feature a specialised and conventionalised use of the term ‘eye’. Ameka distinguishes a psychological and a physical use of eyes. Only the former features in the expression of emotions. It is recognisably different from the physical, body part, eye, which can occur in the plural – after all, we tend to have two eyes. The expressions of negative feelings in the example above can never be used in the plural. There are more differences between ‘ordinary’ and ’emotional’ uses of the word ‘eye’ in Ewe. Read Felix Ameka’s article to which I link below.

But I must go away now and research the origins and exact meaning of the Bambara phrase n sinamuso nyɛ fin ‘ the black eye of my co-wife’. I only know it from the textile design that is named after is, for instance for a wax like the one below. If I remember correctly, the black eye stands for the co-wife’s anger, provoked by her rival wearing such a beautiful wax. When wearing a wrapper bearing the design, the emotion is not evoked through language, but delicately and indirectly through talking fabrics. After all, communicating through speech only would be so limiting!

Wax of the type “N sinamuso nyɛ fin

Here the link to Felix Ameka’s article:

Ameka, Felix K. 2002. Cultural scripting of body parts for emotions: On ‘jealousy’ and related emotions in Ewe. Pragmatics and Cognition 10(1):

Ideas of the sea

Every speaker of a language carries some invisible baggage: how the concepts encoded in the words of their language(s) and their relationships with other words pre-structure the world for them and create expectations on possible connotations and translations. And every learner of a new language knows that a central part of acquiring another language is letting go of these associations in order to fully enter a new language space.

Today, I have two examples that illustrate language worlds and how different they can be in their associations. Both are related to the sea. The first ‘sea’ I’m going to talk about surprised me by being located more than 70 kilometers away from the shores of the Atlantic ocean on the bank of what for me, and in French and English translation, was ‘a river’, ‘un fleuve‘: the Casamance river that reaches from the Atlantic coast more than 320km inland. You can call it a river since it has two banks, a source, and a wide mouth between which is snakes its way through the marsh lands. Or, as Casamançais do, you can call it the sea – jakam in Baïnounk languages for instance. And sea-like this ‘river’ is up to the city of Ziguinchor: till here, it is more like a fjord, vast and filled with salty water carried there by the tides. Its meandering marigots – smaller sea arms – are called cinda in Bainounk Gujaher. Remember cin– the noun class marker derived from the word for ‘rope’ that create words for elongated, rope-like objects? But I digress.

On the Casamance river

The idea of the sea evoked by jakam could not be further removed from the one captured in the word Sahel. The Sahel is a shore, but its sea is a desert: the Sahara. Originating from Arabic sāḥil for ‘shore, coastline’, this word alludes to the sea-like qualities of the desert, whose dunes form waves that can be ridden by travellers. Geographical and sociopolitical borders are often drawn so that the areas north and south of the Sahara are seen as belonging to different spaces. But really, the inhabitants of these regions are also all dwellers of the shore of a sandy ocean that connects them.

(No) strings attached

The observation that many noun class systems – characterised by nominal inflections in which all nouns in a language are formally marked by so-called noun classes or gender – have plants and their classification at their core goes back to Brent Berlin. True to this finding, many Atlantic languages have genders (paired noun classes for singular and one or several plurals) for trees, fruit, seeds and smaller plant-based items. You can find some examples in this post. These noun classes can be extended to items that are similar to botanical items in various respects: they can have a longest vertical axis just like trees, be spherical or have a round diameter like many fruit, or occur in an extended assemblage or a mass just like creepers or beans. Many Baïnounk languages, spoken in and around the Casamance region of Senegal, have an additional gender for string-like objects. Have a look at these nouns that are among those that enter the ‘string’ gender in Baïnounk Gujaher:

Singular form Plural form Gloss
cin-niba ñan-niba ‘shrub of the species Dombeya quinqueseta (Delile) Exell’
ciŋ-ŋaarara ñaŋ-ŋaarara ‘vine of the Smilax anceps Willd.’
cil-lug ñal-lug ‘marrow plant’
ciŋ-kal ñaŋ-kal ‘tail’
cin-díít ñan-díít ‘intestine’
cin-tííb ñan-tííb ‘trace’
cil-líít ñal-líít ‘ribbon’

What is remarkable is the origin of the class marker cin-. It is transparently related to the word denoting ‘bark’, ‘rope’ and ‘string’ – cin-cind. So the root cind occurs with a noun class marker that is probably derived from it to classify rope as a string-like item. The link to the botanical domain is still very salient, as ropes and strings are made from the bark of trees and from the stalks of vines and creepers.

A climbing belt made from plant fiber and string

But what is even neater is the metaphorical extension of the prefix cin– into other domains. It is used to create the noun ‘family name’, cir-ram, literally cin-greet. Greeting is a reciprocal activity that connects two people, and verbal nouns and infinitives linking participants via social activities are created using cin-, in addition to being marked with the reciprocal suffix –ai.

Gujaher infinitive Gloss
cinsukai ‘accompany each other’
cinfeyai ‘hate each other’
cinnannai ‘exchange’ (lit.: cin-give-reciprocal)
cinramai ‘greet each other’
cinyikai ‘have problems with each other’
cimbicai ‘divorce each other’
cimbutai ‘share with each other’
cimmaŋai ‘love each other’

I can hear some of you thinking aloud: “Why is ‘to marry’ not in the list when ‘love’ and ‘divorce’ are? It’s perhaps the most prototypical reciprocal activity!” Well, not in Gujaher (and in fact in many West African languages. While the verbs with cin– above denote activities that are thought about as involving a relationship or an exchange between two equally agentive participants, whose roles can be reversed, ‘marry’ is not construed in this way. The infinitive for ‘marry’ is bujax in Gujaher, taking a different noun class marker to signal the infinitive. The verb can only have men as agents – they are seen as taking women in marriage. Women can only be taken in marriage, so no reciprocity here. They are equal in love, hate and divorce though!

Liquid salt

Most languages of the Niger-Congo stock (a large grouping of languages that are likely to go back to one single ancestor language) have rich noun class systems, on which I’ve written before in several posts. Some families within this large group don’t, among them Mande languages, and this is one of the numerous reasons while the historical relationships between these languages are tenuous. The ways in which noun class systems are organised are really different from language to language, but one striking fact sticks out: the vast majority of Niger-Congo languages with noun classes dedicate a particular noun class to liquids, and, highly unusually, the affix that formally signals this noun class has a similar shape across them. Liquids bear a noun class marker with that has the initial consonant m.

Okay, some of you will say, so if I want to order a coke, I’ll just stick an m in front, and I’ll have a word that could belong to a Niger-Congo language. But you already have a hunch that this would be too easy, don’t you? In noun class systems, items are not put into a noun class based on a single criterion. Rather, these classification systems are built up over time, and while speakers may put an object into a particular class based on different and conflicting criteria (being a ‘natural’ liquid such as sap, blood or tears as opposed to being man-made, for instance) at any point in time, this is even more true over time. But many liquids, and in particular those that occur as liquids in the natural world, are marked with m. Here are the only six words I have found in this class in Baïnounk Gujaher to give you a taste:

Gujaher form Gloss
mun-saal ‘urine’
mun-jil ‘tear’
mun-yin ‘milk’
mul-leen ‘blood’
muŋ-xaana ‘oil’
mum-méér ‘salt

Clearly, all of these words denote liquids. All of them? Some of you may have stumbled over ‘salt’. Why would this crystal be categorised as similar to blood? Looking at how salt is made in Casamance provides the necessary clue. In this area close to the sunken coast line of the Upper Guinea Coast, rivers are tidal and carry salt water far inland. Out of the salty soil close to the rivers, salt is won by extraction through mixing it with water and then evaporating the repeatedly filtered salty water by putting it to the boil until the salt crystals remain. Imagine a future where everybody in Casamance buys salt in shops and this practice is lost. The language would still reflect the experiences of previous generations that salt arrived as water, got absorbed by the soil, and turned into a liquid again before being reduced to its final state.

Here you can see salt and its transformation from solid to liquid and back.

We’re all in it together. But wait, are we really?

Some of you might think this is a post about British politics. But even for those to whom this post does not evoke elusive promises by politicians it might be useful to imagine a language that makes it crystal clear who is meant by a message such as “We’re all in it together.” Does it really intend to include speaker and all addressees, without any wiggle room? For those of you who’d prefer a language that unambiguously signals whether speakers makes an assertion that includes everybody they talk to and about or not, West African languages are here to help.

Many of these languages, for instance Fula, Jalonke, many Baïnounk and Joola languages and Casamance Creole, prove themselves useful by distinguishing in their first person plural pronouns (‘we’) whether the addressees are included (all of us, including YOU) or not (all of us, excluding YOU). In these languages, it makes a crucial difference whether somebody says, as in this example from Baïnounk Gujaher:

Ankëbëndoŋ kahar.
‘We eat meat (including you).’


Ankëbëminiŋ kahar.
‘We (but not you) eat meat.’

If you’ve ever walked away hungry from a dinner table, you’ll get the salience of this difference. Speakers of languages that mark it simply can’t be vague about who is included in the statement, as the languages don’t let them get away with evasiveness in the matter. Other related languages, some of them spoken by populations multilingual in languages with the inclusive/exclusive distinction, do not mark it. Among them are Mandinka, Bambara and Wolof. So clearly, if you’re a politician (or just a random cunning person), and you don’t want to commit to whom your assertion extends to, speak English, Wolof, or Mandinka – but better not Gujaher.

Africans have been here earlier!

From yesterdays’ post on indigenous African languages, an interesting discussion emerged, focussing on differences between languages regarding their lexicalisation patterns. While everything can be expressed in every language, not every language has a single word for every notion. Languages tend to have words for notions which are central to their speakers, and which words these are is influenced by their environment and also shaped by the lexicalisation patterns of other languages present in it. West African languages are noteworthy for their many verbs and derivational affixes that specify particular phases of an event, and also for verbs that denote specifically doing something early. In Bambara (Mande, Mali), these are the verbs soli ‘do something early, do something in the morning’ (note how clunky the English translation equivalent is in comparison) and joona ‘ do something earlier than expected’. In Jalonke (Mande, Guinea), there is the verb kurun ‘leave early’, and the Atlantic language Gujaher (Senegal), also has a verb with the same meaning, fura.

Misty morning in Saare Kindia, in the Futa Jalon region of Guinea